View Content 2

Pacific Earthquake Engineering
Research Center
Capacity-Demand-Diagram Methods for

Estimating Seismic Deformation of Inelastic Structures:
SDF Systems
Anil K. Chopra
University of California, Berkeley
Rakesh K. Goel
California State Polytechnic University
San Luis Obispo
A report on research conducted under grant no. CMS-9812531

from the National Science Foundation: U.S.-Japan Cooperative Research
in Urban Earthquake Disaster Mitigation
PEER 1999/02
APRIL 1999
CAPACITY-DEMAND-DIAGRAM METHODS FOR ESTIMATING
SEISMIC DEFORMATION OF INELASTIC STRUCTURES:
SDF SYSTEMS
Anil K. Chopra
University of California
Berkeley
Rakesh K. Goel
California State Polytechnic University
San Luis Obispo
A report on Research Conducted Under

Grant No. CMS-9812531, U.S.-Japan Cooperative
Research in Urban Earthquake Disaster Mitigation,
from the National Science Foundation
Report No. PEER-1999/02

Pacific Earthquake Engineering Research Center
College of Engineering
University of California, Berkeley
April 1999
ABSTRACT
The ATC-40 and FEMA-274 documents contain simplified nonlinear analysis procedures
to determine the displacement demand imposed on a building expected to deform inelastically.
The Nonlinear Static Procedure in these documents, based on the capacity spectrum method,
involves several approximations: The lateral force distribution for pushover analysis and
conversion of these results to the capacity diagram are based only on the fundamental vibration
mode of the elastic system. The earthquake-induced deformation of an inelastic SDF system is
estimated by an iterative method requiring analysis of a sequence of equivalent linear systems,
thus avoiding the dynamic analysis of the inelastic SDF system. This last approximation is first
evaluated in this report, followed by the development of an improved simplified analysis
procedure, based on capacity and demand diagrams, to estimate the peak deformation of inelastic
SDF systems.
Several deficiencies in ATC-40 Procedure A are demonstrated. This iterative procedure
did not converge for some of the systems analyzed. It converged in many cases, but to a
deformation much different than dynamic (nonlinear response history or inelastic design
spectrum) analysis of the inelastic system. The ATC-40 Procedure B always gives a unique value
of deformation, the same as that determined by Procedure A if it converged.
The peak deformation of inelastic systems determined by ATC-40 procedures are shown
to be inaccurate when compared against results of nonlinear response history analysis and
inelastic design spectrum analysis. The approximate procedure underestimates significantly the
deformation for a wide range of periods and ductility factors with errors approaching 50%,
implying that the estimated deformation is about half the “exact” value.
Surprisingly, the ATC-40 procedure is deficient relative to even the elastic design
spectrum in the velocity-sensitive and displacement-sensitive regions of the spectrum. For
periods in these regions, the peak deformation of an inelastic system can be estimated from the
elastic design spectrum using the well-known equal displacement rule. However, the
approximate procedure requires analyses of several equivalent linear systems and still produces
worse results.
Finally, an improved capacity-demand-diagram method that uses the well-known
constant-ductility design spectrum for the demand diagram has been developed and illustrated by
examples. This method gives the deformation value consistent with the selected inelastic design
spectrum, while retaining the attraction of graphical implementation of the ATC-40 methods.
One version of the improved method is graphically similar to ATC-40 Procedure A whereas a
second version is graphically similar to ATC-40 Procedure B. However, the improved
procedures differ from ATC-40 procedures in one important sense. The demand is determined by
analyzing an inelastic system in the improved procedure instead of equivalent linear systems in
ATC-40 procedures.
The improved method can be conveniently implemented numerically if its graphical
features are not important to the user. Such a procedure, based on equations relating R y and µ
for different T n ranges, has been presented, and illustrated by examples using three different
R y − µ − T n relations.
ii
ACKNOWLEDGMENTS
This research investigation is funded by the National Science Foundation under Grant
CMS-9812531, a part of the U.S.-Japan Cooperative Research in Urban Earthquake Disaster
Mitigation. This financial support is gratefully acknowledged. Dr. Rakesh Goel acknowledges
the State Faculty Support Grant Fellowship received during summer of 1998.
The authors have benefited from discussions with Chris D. Poland and Wayne A. Low,
who are working on a companion research project awarded to Degenkolb Engineers; and with
Sigmund A. Freeman, who, more than any other individual, has been responsible for the concept
and development of the capacity spectrum method.
iii
CONTENTS
ABSTRACT ...................................................................................................................................ii
ACKNOWLEDGMENTS ...........................................................................................................iii
CONTENTS.................................................................................................................................. iv
INTRODUCTION......................................................................................................................... 1
EQUIVALENT LINEAR SYSTEMS.......................................................................................... 3
ATC-40 ANALYSIS PROCEDURES ......................................................................................... 7

PROCEDURE A........................................................................................................................ 8
Examples: Specified Ground Motion...................................................................................... 8
Examples: Design Spectrum ................................................................................................. 15
PROCEDURE B...................................................................................................................... 21
Examples: Specified Ground Motion.................................................................................... 21
Examples: Design Spectrum ................................................................................................. 26
EVALUATION OF ATC-40 PROCEDURES.......................................................................... 29
SPECIFIED GROUND MOTION......................................................................................... 29
DESIGN SPECTRUM ............................................................................................................ 33
IMPROVED PROCEDURES .................................................................................................... 38
INELASTIC DESIGN SPECTRUM ..................................................................................... 38
INELASTIC DEMAND DIAGRAM..................................................................................... 38
PROCEDURE A...................................................................................................................... 40
Examples ............................................................................................................................... 40
Comparison with ATC-40 Procedure A................................................................................ 46
PROCEDURE B...................................................................................................................... 46
Examples ............................................................................................................................... 47
Comparison with ATC-40 Procedure B ................................................................................ 47
ALTERNATIVE DEFINITION OF EQUIVALENT DAMPING ..................................... 47
IMPROVED PROCEDURE: NUMERICAL VERSION........................................................ 51
BASIC CONCEPT .................................................................................................................. 51
RY – µ −TN EQUATIONS ......................................................................................................... 51
CONSISTENT TERMINOLOGY............................................................................................. 55
CONCLUSIONS ......................................................................................................................... 55
NOTATION ................................................................................................................................. 58
REFERENCES ............................................................................................................................ 60
iv
APPENDIX A. DEFORMATION OF VERY-SHORT PERIOD SYSTEMS BY ATC-40
PROCEDURE B.......................................................................................................................... 63
APPENDIX B: EXAMPLES USING R y − µ − T n EQUATIONS ........................................... 65
v
INTRODUCTION
A major challenge to performance-based seismic design and engineering of buildings is

to develop simple, yet effective, methods for designing, analyzing and checking the design of
structures so that they reliably meet the selected performance objectives. Needed are analysis
procedures that are capable of predicting the demands — forces and deformations — imposed by
earthquakes on structures more realistically than has been done in building codes. In response to
this need, simplified, nonlinear analysis procedures have been incorporated in the ATC-40 and
FEMA-274 documents (Applied Technology Council, 1996; FEMA, 1997) to determine the
displacement demand imposed on a building expected to deform inelastically.
The Nonlinear Static Procedure in these documents is based on the capacity spectrum
method originally developed by Freeman et al. (1975) and Freeman (1978). It consists of the
following steps:
1. Develop the relationship between base shear, Vb , and roof (Nth floor) displacement, u N
(Fig. 1a), commonly known as the pushover curve.
2. Convert the pushover curve to a capacity diagram, (Fig. 1b), where
2
N N 
∑ m j φ j1  ∑ m j φ j1 
j =1  j =1  (1)
Γ1 = M1 = N
*
N
∑ m j φ j1 ∑ m j φ j1
2 2
j =1 j =1
and m j = lumped mass at the jth floor level, φ j1 is the jth-floor element of the fundamental
mode φ1 , N is the number of floors, and M 1* is the effective modal mass for the
fundamental vibration mode.
3. Convert the elastic response (or design) spectrum from the standard pseudo-acceleration, A ,
versus natural period, T n , format to the A − D format, where D is the deformation spectrum
ordinate (Fig. 1c).
4. Plot the demand diagram and capacity diagram together and determine the displacement
demand (Fig. 1d). Involved in this step are dynamic analyses of a sequence of equivalent
linear systems with successively updated values of the natural vibration period, T eq , and
equivalent viscous damping, ζ̂ eq (to be defined later).
5. Convert the displacement demand determined in Step 4 to global (roof) displacement and
individual component deformation and compare them to the limiting values for the specified
performance goals.
Approximations are implicit in the various steps of this simplified analysis of an inelastic
MDF system. Implicit in Steps 1 and 2 is a lateral force distribution assumed to be fixed, and
based only on the fundamental vibration mode of the elastic system; however, extensions to
account for higher mode effects have been proposed (Paret et al., 1996). Implicit in Step 4 is the
belief that the earthquake-induced deformation of an inelastic SDF system can be estimated
1
uN Pushover Curve
F
Vb
Vb
uN
(a)
Vb
A=
M1*
Pushover Curve Capacity Diagram
Vb
uN uN
D=
Γ1 φ N1
(b)
Tn,1
2
Tn
D = A
4π
2
A A
Tn,2
Tn D
Demand Diagram
(c)
5% Demand Demand Point

Diagram
Demand Diagram
Capacity Diagram
A
D
(d)
Figure 1. Capacity spectrum method: (a) development of pushover curve, (b) conversion of
pushover curve to capacity diagram, (c) conversion of elastic response spectrum from
standard format to A-D format, and (d) determination of displacement demand.
2
satisfactorily by an iterative method requiring analysis of a sequence of equivalent linear SDF
systems, thus avoiding the dynamic analysis of the inelastic SDF system. This investigation
focuses on the rationale and approximations inherent in this critical step.
The principal objective of this investigation is to develop improved simplified analysis
procedures, based on capacity and demand diagrams, to estimate the peak deformation of
inelastic SDF systems. The need for such procedures is motivated by first evaluating the above
mentioned approximation inherent in Step 4 of the ATC-40 procedure. Thereafter, improved
procedures using the well-established inelastic response (or design) spectrum (e.g., Chopra,
1995; Section 7.10) are developed. The idea of using the inelastic design spectrum in this context
was suggested by Bertero (1995) and introduced by Reinhorn (1997) and Fajfar (1998, 1999);
and the capacity spectrum method has been evaluated previously, e.g., Tsopelas et al. (1997).
EQUIVALENT LINEAR SYSTEMS
The earthquake response of inelastic systems can be estimated by approximate analytical

methods in which the nonlinear system is replaced by an “equivalent” linear system. These
methods attracted the attention of researchers in the 1960s before high speed digital computers
were widely used for nonlinear analyses, and much of the fundamental work was accomplished
over two decades ago (Hudson, 1965; Jennings, 1968; Iwan and Gates, 1979a). In general,
approximate methods for determining the parameters of the equivalent linear system fall into two
categories: methods based on harmonic response and methods based on random response. Six
methods are available in the first category and three in the second category. Formulas for the
natural vibration period and damping ratio are available for each method (Iwan and Gates,
1979a). Generally speaking, the methods based on harmonic response considerably overestimate
the period shift, whereas the methods considering random response give much more realistic
estimates of the period (Iwan and Gates, 1979b).
Now there is renewed interest in applications of equivalent linear systems to design of
inelastic structures. For such applications, the secant stiffness method (Jennings, 1968) is being
used in the capacity spectrum method to check the adequacy of a structural design (e.g., Freeman
et al., 1975; Freeman, 1978; Deierlein and Hsieh, 1990; Reinhorn et al., 1995) and has been
adapted to develop the “nonlinear static procedure” in the ATC-40 report (Applied Technology
Council, 1996) and the FEMA-274 report (FEMA, 1997). A variation of this method, known as
the substitute structure method (Shibata and Sozen, 1976), is popular for displacement-based
design (Gulkan and Sozen, 1974; Shibata and Sozen, 1976; Moehle, 1992; Kowalsky et al.,
1995; Wallace, 1995). Based on harmonic response, these two methods are known to be not as
accurate as methods based on random response (Iwan and Gates, 1979a,b). The equivalent linear
system based on the secant stiffness is reviewed next.
Consider an inelastic SDF system with bilinear force-deformation relationship on initial
loading (Fig. 2a). The stiffness of the elastic branch is k and that of the yielding branch is αk .
The yield strength and yield displacement are denoted by f y and u y , respectively. If the peak
(maximum absolute) deformation of the inelastic system is um , the ductility factor µ = u m u y .
For the bilinear system of Fig. 2a, the natural vibration period of the equivalent linear
system with stiffness equal to k sec , the secant stiffness, is
3
µ (2)
T eq = T n
1 + αµ − α
where T n is the natural vibration period of the system vibrating within its linearly elastic range
( u ≤ u y ).
The most common method for defining equivalent viscous damping is to equate the
energy dissipated in a vibration cycle of the inelastic system and of the equivalent linear system.
Based on this concept, it can be shown that the equivalent viscous damping ratio is (Chopra,
1995: Section 3.9)
1 ED (3)
ζeq =
4π E S
where the energy dissipated in the inelastic system is given by the area E D enclosed by the
hysteresis loop (Fig. 2b) and E S = k sec u 2m / 2 is the strain energy of the system with stiffness k sec
(Fig. 2b). Substituting for E D and E S in Eq. (3) leads to
2 (µ − 1)(1 − α) (4)
ζeq =
π µ(1 + αµ − α)
The total viscous damping of the equivalent linear system is
ζˆ eq = ζ + ζ eq (5)
where ζ is the viscous damping ratio of the bilinear system vibrating within its linearly elastic
range ( u ≤ u y ).
For elastoplastic systems, α = 0 and Eqs. (2) and (4) reduce to

2 µ −1 (6)
T eq = T n µ ζeq =
π µ
Equations (2) and (4) are plotted in Fig. 3 where the variation of T eq T n and ζ eq with µ
is shown for four values of α . For yielding systems ( µ > 1), T eq is longer than T n and ζ eq > 0.
The period of the equivalent linear system increases monotonically with µ for all α . For a fixed
µ , T eq is longest for elastoplastic systems and is shorter for systems with α > 0. For α = 0, ζ eq
increases monotonically with µ but not for α > 0. For the latter case, ζ eq reaches its maximum
value at a µ value, which depends on α, and then decreases gradually.
4
f y (1 + αµ − α )
fy αk
1
Force
ksec
k 1
1
uy um
Deformation
(a)
Force
f y (1 + αµ − α )
fy ES
ED
uy um
Deformation
(b)
Figure 2. Inelastic SDF system: (a) bilinear force-deformation relationship; (b) equivalent
viscous damping due to hysteretic energy dissipation.
5
4
3.5
3
α=0
2.5
0.05
Teq ÷ Tn
2 0.1
0.2
1.5
0.5
0
0 1 2 3 4 5 6 7 8 9 10
µ
(a)
0.7
0.6
α=0
0.5
0.4
ζeq
0.05
0.3
0.1
0.2
0.2
0.1
0
0 1 2 3 4 5 6 7 8 9 10
µ
(b)
Figure 3. Variation of period and viscous damping of the equivalent linear system with
ductility.
6
ATC-40 ANALYSIS PROCEDURES
Contained in the ATC-40 report are approximate analysis procedures to estimate the
earthquake-induced deformation of an inelastic system. These procedures are approximate in the
sense that they avoid dynamic analysis of the inelastic system. Instead dynamic analyses of a
sequence of equivalent linear systems with successively updated values of T eq and ζ̂ eq provide a
basis to estimate the deformation of the inelastic system; T eq is determined by Eq. (2) but ζ̂ eq by
a modified version of Eq. (5):
ζˆ eq = ζ + κ ζ eq (7)
with ζeq limited to 0.45. Although the basis for selecting this upper limit on damping is not
stated explicitly, ATC-40 states that “The committee who developed these damping coefficients
concluded that spectra should not be reduced to this extent at higher values and judgmentally …
set an absolute limit on … [ 0.05+ ζ eq ] of about 50 percent.”
1.2
0.8
Type A
0.6
κ
Type B
0.4
Type C
0.2
0
0 0.1 0.2 0.3 0.4 0.5 0.6
ζ
eq
Figure 4. Variation of damping modification factor with equivalent viscous damping.
The damping modification factor, κ , based primarily on judgment, depends on the

hysteretic behavior of the system, characterized by one of three types: Type A denotes hysteretic
behavior with stable, reasonably full hysteresis loops, whereas Type C represents severely
pinched and/or degraded loops; Type B denotes hysteretic behavior intermediate between Types
A and C. ATC-40 contains equations for κ as a function of ζ eq computed by Eq. (3) for the
three types of hysteretic behavior. These equations, plotted in Fig. 4, were designed to ensure
that κ does not exceed an upper limit, a requirement in addition to the limit of 45% on ζ eq .
7
ATC-40 states that “… they represent the consensus opinion of the product development team.”
Concerned with bilinear systems, this paper will use the κ specified for Type A systems.
ATC-40 specifies three different procedures to estimate the earthquake-induced
deformation demand, all based on the same underlying principles, but differing in
implementation. Procedures A and B are analytical and amenable to computer implementation,
whereas procedure C is graphical and most suited for hand analysis. Designed to be the most
direct application of the methodology, Procedure A is suggested to be the best of the three
procedures. The capacity diagram is assumed to be bilinear in Procedure B. The description of
Procedures A and B that follows is equivalent to that in the ATC-40 report except that it is
specialized for bilinear systems.
PROCEDURE A
This procedure in the ATC-40 report is described herein as a sequence of steps:

1. Plot the force-deformation diagram and the 5%-damped elastic response (or design) diagram,
both in the A − D format to obtain the capacity diagram and 5%-damped elastic demand
diagram, respectively.
2. Estimate the peak deformation demand Di and determine the corresponding pseudo-
acceleration Ai from the capacity diagram. Initially, assume Di = D(T n , ζ = 5%) , determined
for period T n from the elastic demand diagram.
3. Compute ductility µ = Di ÷ u y .
4. Compute the equivalent damping ratio ζ̂ eq from Eq. (7).
5. Plot the elastic demand diagram for ζ̂ eq determined in Step 4 and read-off the displacement
D j where this diagram intersects the capacity diagram.
6. Check for convergence. If ( D j − Di ) ÷ D j ≤ tolerance (=0.05) then the earthquake induced
deformation demand D = D j . Otherwise, set Di = D j (or another estimated value) and repeat
Steps 3-6.
Examples: Specified Ground Motion

This procedure is used to compute the earthquake-induced deformation of the six
example systems listed in Table 1. Considered are two values of T n : 0.5s in the acceleration-
sensitive spectral region and 1s in the velocity-sensitive region, and three levels of yield strength;
ζ=5% for all systems. The excitation chosen is the north-south component of the El Centro
ground motion; the particular version used is from Chopra (1995). Implementation details are
presented next for selected systems and final results for all systems in Table 2.
The procedure is implemented for System 5 (Table 1).
1. Implementation of Step 1 gives the 5%-damped elastic demand diagram and the capacity
diagram in Fig. 5a.
8
2. Assume Di = D(1.0,5%) = 11.27 cm .
3. µ = 11.27 ÷ 2.562 = 4.40 .
4. ζ eq = 0.637 × (4.40 − 1) ÷ 4.40 = 0.49 ; instead use the maximum allowable value 0.45. For
ζ eq = 0.45 and Type A systems (Fig. 4), κ = 0.77 and ζˆ = ζ + κ ζ eq = 0.05 + 0.77 × 0.45
eq
= 0.397 .
5. The elastic demand diagram for 39.7% damping intersects the capacity diagram at
D j = 3.725 cm (Fig. 5a).
6. 100 × ( D j − Di ) ÷ D j = 100 × (3.725 − 11.27 ) ÷ 3.725 = −202.6% > 5% tolerance. Set
Di = 3.725 cm and repeat Steps 3 to 6.
For the second iteration, Di = 3.725 cm , µ = 3.725 ÷ 2.562 = 1.45 ,
ζ eq = 0.637 × (1.45 − 1) ÷ 1.45 = 0.198 , κ= 0.98, and ζˆ eq = 0.243 . The intersection point
D j = 5.654 cm and the difference between Di and D j = 34.1% which is greater than the 5%
tolerance. Therefore, additional iterations are required; results of these iterations are summarized
in Table 3. Error becomes less than the 5% tolerance at the end of sixth iteration and the
procedure could have been stopped there. However, the procedure was continued until the error
became practically equal to zero. The deformation demand at the end of the iteration process,
D j = 4.458 cm . Determined by response history analysis (RHA) of the inelastic system,
D exact = 10.16 cm and the error = 100 × (4.458 − 10.16 ) ÷ 10.16 = -56.1%.
Fig. 5b shows the convergence behavior of the ATC-40 Procedure A for System 5.
Observe that this iterative procedure converges to a deformation much smaller than the exact
value. Thus convergence here is deceptive because it can leave the erroneous impression that the
calculated deformation is accurate. In contrast, a rational iterative procedure should lead to the
exact result after a sufficient number of iterations.
The procedure is next implemented for System 6 (Table 1).
1. Implementation of Step 1 gives the 5%-damped elastic demand diagram and capacity
diagram in Fig. 6a.
2. Assume Di = D(1.0,5%) = 11.27 cm .
3. µ = 11.27 ÷ 4.302 = 2.62 .
4. ζ eq = 0.637 × (2.62 − 1) ÷ 2.62 = 0.39 . For ζ eq = 0.39 and Type A systems (Fig. 4),
κ = 0.82 and ζˆ = ζ + κ ζ eq = 0.05 + 0.82 × 0.39 = 0.371 .
eq
D j = 3.538 cm (Fig. 6a).
9
The results for subsequent iterations, summarized in Table 4, indicate that the procedure
fails to converge for this example. In the first iteration, the 37.1%-damped elastic demand
diagram intersects the capacity diagram in its linear-elastic region (Fig. 6a). In subsequent
iterations, the intersection point alternates between 11.73 cm and 3.515 cm (Table 4 and Fig. 6b).
In order to examine if the procedure would converge with a new starting point, the procedure
was restarted with Di = 5 cm at iteration number 7. However, the procedure diverges very
quickly as shown by iterations 7 to 15 (Table 4 and Fig. 6b), ending in an alternating pattern.
Table 1. Properties of example systems and their response to El Centro (1940) ground
motion.
System Properties System Response

System Tn f y÷w uy µ D exact
(s) (cm) (cm)
1 0.5 0.1257 0.7801 6 4.654
2 0.1783 1.106 4 4.402
3 0.3411 2.117 2 4.210
4 1 0.07141 1.773 6 10.55
5 0.1032 2.562 4 10.16
6 0.1733 4.302 2 8.533
Table 2. Results from ATC-40 Procedure A analysis of six systems for El Centro (1940)
ground motion.
System Converged D approx D exact Error

(?) (cm) (cm) (%)
1 Yes 3.534 4.654 -24.1
2 Yes 3.072 4.402 -30.2
3 No -- 4.210 --
4 Yes 7.912 10.55 -25.0
5 Yes 4.458 10.16 -56.1
6 No -- 8.533 --
10
Table 3. Detailed results from ATC-40 Procedure A analysis of System 5 for El Centro
(1940) ground motion.
Iteration Di Ai ζˆ eq Dj Aj Difference
No. (%)
1 11.272 0.1032 0.3965 3.7252 0.1032 -202.6
2 3.7252 0.1032 0.2432 5.6537 0.1032 34.1
3 5.6537 0.1032 0.3466 4.0832 0.1032 -38.5
4 4.0832 0.1032 0.2732 4.7214 0.1032 13.5
5 4.7214 0.1032 0.3114 4.3523 0.1032 -8.5
6 4.3523 0.1032 0.2912 4.5002 0.1032 3.3
7 4.5002 0.1032 0.2999 4.4425 0.1032 -1.3
8 4.4425 0.1032 0.2966 4.4639 0.1032 0.5
9 4.4639 0.1032 0.2978 4.4561 0.1032 -0.2
10 4.4561 0.1032 0.2974 4.4589 0.1032 0.1
11 4.4589 0.1032 0.2975 4.4579 0.1032 0
11
0.5 s 1s
1
EC
0.8
5% Demand Diagram
0.6
A, g
2s
(11.27,0.4541)
0.4 29.7% Demand Diagram (Last Iteration)
39.7% Demand Diagram (1st Iteration)
0.2 (3.725,0.1032)
(4.458,0.1032) Capacity Diagram
(11.27,0.1032) 5s
10 s
0
0 10 20 30 40 50
D, cm
(a)
12
EC
Exact
10
8
Dj, cm
0
0 2 4 6 8 10 12
Iteration No.
(b)
Figure 5. Application of ATC-40 Procedure A to System 5 for El Centro (1940) ground

motion: (a) iterative procedure, and (b) convergence behavior.
12
Table 4. Detailed results from ATC-40 Procedure A analysis System 6 for El Centro (1940)
ground motion.
Iteration Di Ai ζ̂ eq Dj Aj Difference Comments

No. (%)
1 11.272 0.1733 0.371 3.5376 0.14251 -218.6 Start with elastic
response
2 3.5376 0.1733 0.05 11.726 0.1733 69.8
3 11.726 0.1733 0.3756 3.5146 0.14158 -233.6
4 3.5146 0.1733 0.05 11.726 0.1733 70
5 11.726 0.1733 0.3756 3.5146 0.14158 -233.6
6 3.5146 0.1733 0.05 11.726 0.1733 70 Indefinite oscillation
7 5 0.1733 0.1389 5.6491 0.1733 11.5 Iteration restarted
8 5.6491 0.1733 0.2019 4.856 0.1733 -16.3
9 4.856 0.1733 0.1227 6.1903 0.1733 21.6
10 6.1903 0.1733 0.2394 4.2884 0.17276 -44.3
11 4.2884 0.1733 0.05 11.726 0.1733 63.4
12 11.726 0.1733 0.3756 3.5146 0.14158 -233.6
13 3.5146 0.1733 0.05 11.726 0.1733 70
14 11.726 0.1733 0.3756 3.5146 0.14158 -233.6
15 3.5146 0.1733 0.05 11.726 0.1733 70 Failure to converge
16 5.3 0.1733 0.17 5.3711 0.1733 1.3 Iteration restarted
17 5.3711 0.1733 0.1768 5.2752 0.1733 -1.8
18 5.2752 0.1733 0.1675 5.4011 0.1733 2.3
19 5.4011 0.1733 0.1796 5.2299 0.1733 -3.3
20 5.2299 0.1733 0.163 5.4519 0.1733 4.1
21 5.4519 0.1733 0.1844 5.1522 0.1733 -5.8
22 5.1522 0.1733 0.1551 5.5307 0.1733 6.8
23 5.5307 0.1733 0.1915 5.0337 0.1733 -9.9
24 5.0337 0.1733 0.1426 5.6263 0.1733 10.5
25 5.6263 0.1733 0.1999 4.8902 0.1733 -15.1
26 4.8902 0.1733 0.1266 6.0294 0.1733 18.9
27 6.0294 0.1733 0.2296 4.3819 0.1733 -37.6
28 4.3819 0.1733 0.0616 11.077 0.1733 60.4
29 11.077 0.1733 0.3688 3.5484 0.14294 -212.2
30 3.5484 0.1733 0.05 11.726 0.1733 69.7
31 11.726 0.1733 0.3756 3.5146 0.14158 -233.6
32 3.5146 0.1733 0.05 11.726 0.1733 70
33 11.726 0.1733 0.3756 3.5146 0.14158 -233.6 Slow divergence
13
0.5 s 1s
1
EC
0.8
5% Demand Diagram
0.6
A, g
2s
0.4
0.2 (11.73,0.1733) Iterations 2 & 4

(0.1425,3.538) Iteration 1 Capacity Diagram
(0.1416,3.515) Iteration 3 5s
10 s
0
0 10 20 30 40 50
D, cm
(a)
12
10
8
Dj, cm
2
EC
0
0 3 6 9 12 15 18 21 24 27 30 33
Iteration No.
(b)
Figure 6. Application of ATC-40 Procedure A to System 6 for El Centro (1940) ground

motion: (a) iterative procedure, and (b) convergence behavior.
14
Examples: Design Spectrum
The ATC-40 Procedure A is next implemented to analyze systems with the excitation
specified by a design spectrum. For illustration we have selected the design spectrum of Fig. 7,
which is the median-plus-one-standard-deviation spectrum constructed by the procedures of
Newmark and Hall (1982), as described in Chopra (1995; Section 6.9). The systems analyzed
have the same T n as those considered previously but their yield strengths for the selected µ
values were determined from the design spectrum (Table 5). Implementation details are
presented next for selected systems and the final results for all systems in Table 6.
10
Acceleration Velocity Displacement
5 Sensitive Sensitive Sensitive
b c
2
a
1
0.5 d
0.2
A, g
Tf = 33 sec
0.1 e
0.05 T = 1/33 sec
a
0.02 T T
c d
0.01 Tb = 0.125 sec Te = 10 sec
0.005
f
0.002
0.02 0.05 0.1 0.2 0.5 1 2 5 10 20 50
Tn, s
Figure 7. Newmark-Hall elastic design spectrum.
The procedure is implemented for System 5 (Table 5).

diagram in Fig. 8a.
2. Assume Di = D(1.0,5%) = 44.64 cm .
3. µ = 44.64 ÷ 11.16 = 4 .
4. ζeq = 0.637 × (4.0 − 1) ÷ 4.0 = 0.48 ; instead use the maximum allowable value 0.45. For
ζ eq = 0.45 and Type A systems (Fig. 4), κ = 0.77 and ζˆ = ζ + κ ζ eq = 0.05 + 0.77 × 0.45
eq
= 0.397 .
15
D j = 28.18 cm (Fig. 8a).
6. 100 × ( D j − Di ) ÷ D j = 100 × (28.18 − 44.64 ) ÷ 28.18 = −58.4% >5% tolerance. Set Di = 28.18
cm and repeat Steps 3 to 6.
For the second iteration, Di = 28.18 cm , µ = 28.18 ÷ 11.16 = 2.52 ,
ζeq = 0.637 × (2.52 − 1) ÷ 2.52 = 0.38 , κ= 0.84, and ζˆ eq = 0.37 . The intersection point
D j = 31.55 cm and the difference between Di and D j = 10.7% which is greater than the 5%
tolerance. Therefore, additional iterations are required; results of these iterations are summarized
in Table 7. The error becomes less than the 5% tolerance at the end of fourth iteration and the
procedure could have been stopped there. However, the procedure was continued till the error
became practically equal to zero. The deformation demand at the end of the iteration process is
D j = 30.44 cm .
Determined directly from the inelastic design spectrum, constructed by the procedures of
Newmark and Hall (1982), as described in Chopra (1995, Section 7.10), the “reference” value of
deformation is D spectrum = 44.64 cm and the discrepancy = 100 × (30.44 − 44.64 ) ÷ 44.64 = -31.8%.
Fig. 8b shows the convergence behavior of the ATC-40 Procedure A for System 5.
Observe that the iterative procedure converges to a deformation value much smaller than the
“reference” value.
The procedure is next implemented for System 6 (Table 5).
diagram in Fig. 9a.
2. Assume Di = D(0.5,5%) = 44.64 cm .
3. µ = 44.64 ÷ 22.32 = 2.0 .
4. ζeq = 0.637 × (2.0 − 1) ÷ 2.0 = 0.32 . For ζ eq = 0.32 and Type A systems (Fig. 4), κ = 0.87 and
ζˆ = ζ + κ ζeq = 0.05 + 0.87 × 0.32 = 0.33 .
eq
5. The elastic demand diagram for 33% damping intersects the capacity diagram at
D j = 18.56 cm (Fig. 9a).
The results for subsequent iterations, summarized in Table 8, indicate that the procedure
fails to converge for this example. In the first iteration, the 33%-damped elastic demand diagram
intersects the capacity diagram in its linear-elastic region (Fig. 9a). In subsequent iterations, the
intersection point alternates between 13.72 cm and 89.28 cm (Table 8 and Fig. 9b). In order to
examine if the procedure would converge with a new starting point, the procedure was restarted
with Di = 28 cm at iteration number 6. However, the procedure diverges very quickly as shown
by iterations 6 to 11 (Table 8 and Fig. 9b), ending in an alternating pattern.
16
Table 5. Properties of example systems and their deformations from inelastic design
spectrum.
System Properties System Response

System Tn f y÷w uy µ D spectrum
(s) (cm) (cm)
1 0.5 0.5995 3.7202 6 22.32
2 0.8992 5.5803 4 22.32
3 1.5624 9.6962 2 19.39
4 1 0.2997 7.4403 6 44.64
5 0.4496 11.160 4 44.64
6 0.8992 22.321 2 44.64
Table 6. Results from ATC-40 Procedure A analysis of six systems for design spectrum.
System Converged D approx D spectrum Discrepancy

(?) (cm) (cm) (%)
1 No -- 22.32 --
2 No -- 22.32 --
3 No -- 19.39 --
4 No -- 44.64 --
5 Yes 30.44 44.64 -31.8
6 Yes 42.28 44.64 -5.3
Table 7. Detailed results ATC-40 Procedure A analysis of System 5 for design spectrum.
Iteration Di Ai ζˆ eq Dj Aj Difference
No. (%)
1 44.64 0.4496 0.3965 28.18 0.4496 -58.4
2 28.18 0.4496 0.3664 31.55 0.4496 10.7
3 31.54 0.4496 0.3796 30.01 0.4496 -5.1
4 30.01 0.4496 0.3741 30.64 0.4496 2
5 30.64 0.4496 0.3764 30.37 0.4496 -0.9
6 30.36 0.4496 0.3754 30.48 0.4496 0.4
7 30.48 0.4496 0.3759 30.43 0.4496 -0.2
8 30.43 0.4496 0.3757 30.45 0.4496 0.1
9 30.45 0.4496 0.3757 30.44 0.4496 0
17
0.5 s 1s
3
NH
5% Demand Diagram
2 2s
(44.64,1.798)
A, g
39.7% Demand Diagram (1st Iteration)

37.6% Demand Diagram (Last Iteration)
(28.18,0.4492)
1 (30.44,0.4492)
Capacity Diagram
(44.64,0.4492) 5s
10 s
0
0 50 100 150 200
D, cm
(a)
50
Inelastic Design Spectrum
40
30
D , cm
j
20
10
NH
0
0 2 4 6 8 10
Iteration No.
(b)
Figure 8. Application of ATC-40 Procedure A to Example 5 for elastic design spectrum: (a)
iterative procedure, and (b) convergence behavior.
18
Table 8. Detailed results ATC-40 Procedure A analysis of System 6 for design spectrum.
Iteration Di Ai ζ̂ eq Dj Aj Difference Comments

No. (%)
1 44.64 0.8992 0.3288 18.56 0.7475 -140.6 Start with elastic response
2 18.56 0.8992 0.0500 89.28 0.8992 79.2
3 89.28 0.8992 0.3965 13.72 0.5527 -550.7
4 13.72 0.8992 0.0500 89.28 0.8992 84.6
5 89.28 0.8992 0.3965 13.72 0.5527 -550.7 Indefinite oscillation
7 35.27 0.8992 0.2705 23.10 0.8992 -52.7
8 23.10 0.8992 0.0715 71.67 0.8992 67.8
9 71.67 0.8992 0.3917 14.03 0.5653 -410.7
10 14.03 0.8992 0.0500 89.28 0.8992 84.3
11 89.28 0.8992 0.3965 13.72 0.5527 -550.7 Indefinite oscillation
13 32.29 0.8992 0.2413 26.23 0.8992 -23.1
14 26.23 0.8992 0.1449 42.56 0.8992 38.4
15 42.56 0.8992 0.3189 19.34 0.7791 -120.1
16 19.34 0.8992 0.0500 89.28 0.8992 78.3
17 89.28 0.8992 0.3965 13.72 0.5527 -550.7
18 13.72 0.8992 0.0500 89.28 0.8992 84.6
19 89.28 0.8992 0.3965 13.72 0.5527 -550.7 Slow divergence
19
0.5 s 1s
3
NH
5% Demand Diagram
A, g 2 2s
(89.28,0.8992) Iterations 2 & 4

1
(18.56,0.7475) Iteration 1 Capacity Diagram
(13.72,0.5527) Iterations 3 & 5
5s
10 s
0
0 50 100 150 200
D, cm
(a)
100
80
60
Dj, cm
40
20
NH
0
0 2 4 6 8 10 12 14 16 18 20
Iteration No.
(b)
Figure 9. Application of ATC-40 Procedure A to System 6 for elastic design spectrum: (a)
iterative procedure, and (b) convergence behavior.
20
PROCEDURE B
This procedure in the ATC-40 report is described herein as a sequence of steps:

1. Plot the capacity diagram.
2. Estimate the peak deformation demand Di . Initially assume Di = D(T n , ζ = 5%) .
3. Compute ductility µ = Di ÷ u y .
4. Compute equivalent period T eq and damping ratio ζ̂ eq from Eqs. (2) and (7), respectively.
5. Compute the peak deformation D(T eq , ζˆ eq ) and peak pseudo-acceleration A(T eq , ζˆ eq ) of an

elastic SDF system with vibration properties T eq and ζ̂ eq .
6. Plot the point with coordinates D(T eq , ζˆ eq ) and A(T eq , ζˆ eq ) .
7. Check if the curve generated by connecting the point plotted in Step 6 to previously
determined, similar points intersects the capacity diagram. If not, repeat Steps 3-7 with a new
value of Di ; otherwise go to Step 8.
8. The earthquake-induced deformation demand is given by the D -value at the intersection
point.
Examples: Specified Ground Motion

Procedure B is implemented for the Systems 1 to 6 (Table 1). The final results are
summarized in Table 9; details are presented next. For a number of assumed values of µ (or D),
pairs of values D(T eq , ζˆ eq ) and A(T eq , ζˆ eq ) are generated (Tables 10 and 11). These pairs are
plotted to obtain the curve A-B in Fig. 10, wherein capacity diagrams for three systems are
shown together with the 5%-damped linear elastic demand diagram; the latter need not be
plotted. The intersection point between the curve A-B and the capacity diagram of a system gives
its deformation demand: D = 3.536 cm , D = 3.075 cm , and D = 3.284 cm for Systems 1 to 3,
respectively (Fig. 10a); and D = 7.922 cm , D = 4.454 cm , and D = 5.318 cm for Systems 4 to
6, respectively (Fig. 10b). In contrast, the exact deformations computed by RHA of the inelastic
systems are 4.654 cm, 4.402 cm, and 4.210 cm for Systems 1 to 3; and 10.55 cm, 10.16 cm, and
8.533 cm for Systems 4 to 6, indicating that the error in the approximate procedure ranges from –
22% to –56.2%. Observe that the curve A-B provides the information to determine the
deformation demand in several systems with the same T n values but different yield strengths.
Procedure B always gives a unique estimate of the deformation, whereas, as noted earlier,
the iterative Procedure A may not always converge. If it does converge, the two procedures give
the same value of deformation (within round-off and interpolation errors) in the examples solved.
21
Table 9. Results from ATC-40 Procedure B analysis of six systems for El Centro ground
motion.
System D approx D exact Discrepancy

(cm) (cm) (%)
1 3.536 4.654 -24.0
2 3.075 4.402 -30.1
3 3.284 4.210 -22.0
4 7.922 10.55 -24.9
5 4.453 10.16 -56.2
6 5.318 8.533 -37.7
22
Table 10. Detailed results from ATC-40 Procedure B analysis of Systems 1 to 3 for El
Centro (1940) ground motion.
µ T eq ζˆ eq D A
1 0.5 0.05 5.6846 0.91599
1.1 0.5244 0.1079 4.411 0.64616
1.2 0.5477 0.1562 4.0093 0.53837
1.3 0.5701 0.197 3.7297 0.46229
1.4 0.5916 0.2292 3.5032 0.40321
1.5 0.6124 0.2539 3.3449 0.35933
1.55 0.6225 0.2646 3.2859 0.3416
1.5515 0.6228 0.2649 3.2842 0.3411
1.5531 0.6228 0.2649 3.2854 0.34085
1.6 0.6325 0.2743 3.228 0.3251
1.7 0.6519 0.2914 3.1162 0.29537
1.8 0.6708 0.3058 3.0093 0.26939
1.9 0.6892 0.3182 2.9207 0.2477
2 0.7071 0.3288 2.8739 0.23155
2.1 0.7246 0.338 2.898 0.22237
2.2 0.7416 0.3461 2.9204 0.2139
2.3 0.7583 0.3532 2.9412 0.20606
2.4 0.7746 0.3595 2.9605 0.19877
2.5 0.7906 0.365 2.9929 0.19291
2.6 0.8062 0.37 3.024 0.18742
2.7 0.8216 0.3745 3.0531 0.18221
2.7787 0.8335 0.3778 3.0747 0.1783
2.8 0.8367 0.3786 3.0804 0.17727
2.9 0.8515 0.3823 3.1058 0.17257
3 0.866 0.3856 3.1295 0.16809
3.1 0.8803 0.3887 3.1517 0.16382
3.2 0.8944 0.3914 3.1723 0.15974
3.3 0.9083 0.394 3.1992 0.15621
3.4 0.922 0.3964 3.2273 0.15295
3.5 0.9354 0.3965 3.2632 0.15023
3.6 0.9487 0.3965 3.2973 0.14759
3.7 0.9618 0.3965 3.3293 0.14499
3.8 0.9747 0.3965 3.359 0.14244
3.9 0.9874 0.3965 3.3868 0.13993
4 1 0.3965 3.4126 0.13747
4.1 1.0124 0.3965 3.4366 0.13506
4.2 1.0247 0.3965 3.4588 0.1327
4.3 1.0368 0.3965 3.4792 0.13038
4.4 1.0488 0.3965 3.5023 0.12826
4.5 1.0607 0.3965 3.528 0.12633
4.525 1.0636 0.3965 3.5342 0.12585
4.533 1.0645 0.3965 3.5362 0.1257
23
Table 11. Detailed results from ATC-40 Procedure B analysis of Systems 4 to 6 for El
Centro (1940) ground motion.
µ T eq ζˆ eq D A
1 1 0.05 11.272 0.45407
1.1 1.0488 0.1079 7.4536 0.27297
1.2 1.0954 0.1562 5.7066 0.19157
1.225 1.1068 0.167 5.4288 0.17852
1.2362 1.1119 0.1717 5.3182 0.17330
1.2375 1.1119 0.1717 5.3157 0.17304
1.3 1.1402 0.197 5.0373 0.15609
1.4 1.1832 0.2292 4.8741 0.14025
1.5 1.2247 0.2539 4.7401 0.12730
1.6 1.2649 0.2743 4.6154 0.11620
1.7 1.3038 0.2914 4.4969 0.10656
1.7384 1.3185 0.2972 4.4535 0.1032
1.8 1.3416 0.3058 4.3893 0.09823
1.9 1.3784 0.3182 4.5607 0.09670
2 1.4142 0.3288 4.7224 0.09512
2.1 1.4491 0.338 4.8831 0.09367
2.2 1.4832 0.3461 5.0397 0.09228
2.3 1.5166 0.3532 5.1901 0.09090
2.4 1.5492 0.3595 5.335 0.08955
2.5 1.5811 0.365 5.4748 0.08822
2.6 1.6125 0.37 5.6097 0.08692
2.7 1.6432 0.3745 5.7401 0.08564
2.8 1.6733 0.3786 5.8749 0.08452
2.9 1.7029 0.3823 6.0057 0.08343
3 1.7321 0.3856 6.1323 0.08234
3.1 1.7607 0.3887 6.2546 0.08128
3.2 1.7889 0.3914 6.3727 0.08023
3.3 1.8166 0.394 6.491 0.07924
3.4 1.8439 0.3964 6.6094 0.07831
3.5 1.8708 0.3965 6.742 0.07760
3.6 1.8974 0.3965 6.8714 0.07689
3.7 1.9235 0.3965 7.0005 0.07622
3.8 1.9494 0.3965 7.1297 0.07558
3.9 1.9748 0.3965 7.2543 0.07493
4 2 0.3965 7.3773 0.07430
4.1 2.0248 0.3965 7.501 0.07370
4.2 2.0494 0.3965 7.6197 0.07308
4.3 2.0736 0.3965 7.7332 0.07245
4.4 2.0976 0.3965 7.8442 0.07182
4.45 2.1095 0.3965 7.9008 0.07152
4.4688 2.114 0.3965 7.9217 0.07141
24
0.5 s 1s
1
A EC
0.8
5% Demand Diagram
0.6
A, g
2s
0.4 (3.284,0.3411)
0.2 (3.075,0.1783)
(3.536,0.1257)
B 5s
10 s
0
0 10 20 30 40 50
D, cm
(a)
0.5 s 1s
1
EC
0.8
5% Demand Diagram
0.6
A, g
2s
A
0.4
0.2 (5.318,0.1733)
(4.454,0.1032)
5s
B (7.922,0.07141) 10 s
0
0 10 20 30 40 50
D, cm
(b)
Figure 10. Application of ATC-40 Procedure B for El Centro (1940) ground motion: (a)
Systems 1 to 3, and (b) Systems 4 to 6.
25
Examples: Design Spectrum
Procedure B is implemented for the Systems 1 to 6 (Table 5). The results from this
procedure are summarized in Table 12 and illustrated in Fig. 11 where the estimated
deformations are noted; intermediate results are available in Tables 13 and 14. These
approximate values are compared in Table 12 against the values determined directly from the
inelastic design spectrum constructed by the procedure of Newmark and Hall (1982), as
described in Chopra (1995, Section 7.10); see Appendix B for details. Relative to these reference
values, the discrepancy ranges from –5.2% to –58.6% for the systems considered.
Table 12. Results from ATC-40 Procedure B analysis of six systems for design spectrum.
System D approx D spectrum Discrepancy

(cm) (cm) (%)
1 10.46 22.32 -53.1
2 9.245 22.32 -58.6
3 11.51 19.39 -40.6
4 42.27 44.64 -5.2
5 30.45 44.64 -31.7
6 29.84 44.64 -33.1
Table 13. Detailed results from ATC-40 Procedure B analysis of Systems 1 to 3 for design
spectrum.
µ T eq ζˆ eq D A
1 0.5 0.05 16.794 2.7062
1.1 0.5244 0.1079 13.012 1.9061
1.2 0.5477 0.1562 11.332 1.5217
1.1871 0.5448 0.1504 11.51 1.5624
1.3 0.5701 0.197 10.328 1.2802
1.4 0.5916 0.2292 9.7547 1.1227
1.5 0.6124 0.2539 9.4602 1.0163
1.6 0.6325 0.2743 9.2924 0.9358
1.6567 0.6436 0.2843 9.2447 0.8992
1.7 0.6519 0.2914 9.2107 0.8730
1.8 0.6708 0.3058 9.1906 0.8227
1.9 0.6892 0.3182 9.2164 0.7816
2 0.7071 0.3288 9.2776 0.7475
2.1 0.7246 0.338 9.3664 0.7187
2.2 0.7416 0.3461 9.4776 0.6942
2.3 0.7583 0.3532 9.607 0.6731
2.4 0.7746 0.3595 9.7515 0.6547
2.5 0.7906 0.365 9.9087 0.6387
2.6 0.8062 0.37 10.077 0.6245
2.7 0.8216 0.3745 10.254 0.6120
2.8 0.8367 0.3786 10.439 0.6008
2.8123 0.8385 0.3791 10.463 0.5995
26
Table 14. Detailed results from ATC-40 Procedure B analysis of Systems 4 to 6 for design
spectrum.
µ T eq ζˆ eq D A
1 1 0.05 44.642 1.7984
1.2 1.0954 0.1562 32.69 1.0974
1.3519 1.1627 0.2154 29.84 0.8892
1.4 1.1832 0.2292 29.411 0.8463
1.6 1.2649 0.2743 28.488 0.7173
1.8 1.3416 0.3058 28.32 0.6338
2 1.4142 0.3288 28.521 0.5745
2.2 1.4832 0.3461 28.926 0.5297
2.4 1.5492 0.3595 29.448 0.4943
2.6 1.6125 0.37 30.042 0.4655
2.7283 1.6517 0.3757 30.449 0.4496
2.8 1.6733 0.3786 30.679 0.4414
3 1.7321 0.3856 31.343 0.4209
3.2 1.7889 0.3914 32.022 0.4031
3.4 1.8439 0.3964 32.708 0.3875
3.6 1.8974 0.3965 33.648 0.3765
3.8 1.9494 0.3965 34.57 0.3665
4 2 0.3965 35.468 0.3572
4.2 2.0494 0.3965 36.344 0.3486
4.4 2.0976 0.3965 37.199 0.3406
4.6 2.1448 0.3965 38.035 0.3331
4.8 2.1909 0.3965 38.853 0.3261
5 2.2361 0.3965 39.654 0.3195
5.2 2.2804 0.3965 40.44 0.3133
5.4 2.3238 0.3965 41.21 0.3074
5.6 2.3664 0.3965 41.966 0.3019
5.6822 2.3837 0.3965 42.273 0.2997
27
0.5 s 1s
3
A NH
5% Elastic Demand Diagram
2 2s
(11.51,1.56)
A, g
1 (9.245,0.8992)
(10.46,0.5995)
B
5s
10 s
0
0 50 100 150 200
D, cm
(a)
0.5 s 1s
3
NH
2 2s
A
A, g
1 (29.84,0.8992)
(30.45,0.4496)
5s
B (42.27,0.2997)
10 s
0
0 50 100 150 200
D, cm
(b)
Figure 11. Application of ATC-40 Procedure B for elastic design spectrum: (a) Systems 1 to
3, and (b) Systems 4 to 6.
28
EVALUATION OF ATC-40 PROCEDURES
SPECIFIED GROUND MOTION
The ATC-40 Procedure B is implemented for a wide range of system parameters and
excitations in two versions: (1) κ = 1, i.e., the equivalent viscous damping is given by Eqs. (4)
and (5) based on well-established principles; and (2) κ is given by Fig. 4, a definition based
primarily on judgment to account for different types of hysteretic behavior.
The yield strength of each elastoplastic system analyzed was chosen corresponding to an
allowable ductility µ :
f y = ( Ay g )w (8)
where w is the weight of the system and A y is the pseudo-acceleration corresponding to the
allowable ductility and the vibration properties — natural period T n and damping ratio ζ — of
the system in its linear range of vibration. Recall that the ductility demand (computed by
nonlinear response history analysis) imposed by the selected ground motion on systems defined
in this manner will exactly equal the allowable ductility (Chopra, 1995; Section 19.1.1).
The peak deformation due to a selected ground motion, determined by the ATC-40
method, D approx , is compared in Fig. 12 against the “exact” value, D exact , determined by
nonlinear RHA, and the percentage error in the approximate result is plotted in Fig. 13. These
figures permit several observations. The approximate procedure is not especially accurate. It
underestimates significantly the deformation for wide ranges of T n values with errors
approaching 50%, implying that the estimated deformation is only about half of the value
determined by nonlinear RHA. The approximate method gives larger deformation for short
period systems ( T n < 0.1 sec for µ = 2 and T n < 0.4 sec for µ = 6) and the deformation does not
approach zero as T n goes to zero. This unreasonable discrepancy occurs because, for very short-
period systems with small yield strength, the T eq has to shift to the constant-V region of the
spectrum before the capacity and demand diagrams can intersect (Appendix A). While inclusion
of the damping modification factor κ increases the estimated displacement, the accuracy of the
approximate results improves only marginally for the smaller values of µ. Therefore the κ factor
is not attractive, especially because it is based primarily on judgement.
Shown in Fig. 14 are the errors in the ATC-40 method, with the κ factor included, for six
different ground motions: (1) El Centro, S00E, 1940 Imperial Valley; (2) Corralitos, Chan-1, 90
deg, 1989 Loma Prieta; (3) Sylmar County Hospital Parking Lot, Chan-3, 360 deg, 1994
Northridge; (4) Pacoima Dam, N76W, 1971 San Fernando; (5) Lucerne Valley, S80W, 1992
Landers; and (6) SCT, S00E, 1985 Mexico City. Observe that, contrary to intuition, the error
does not decrease consistently for smaller ductility. While the magnitude of the error and its
variation with T n depend on the excitation, the earlier observation that the error in the
approximate method is significant is supported by results for several ground motions.
29
30
µ=2
25
Exact
ATC−40, κ=1
ATC−40
20
D, cm
15
10
5
EC
0
0 0.5 1 1.5 2 2.5 3
Tn, s
(a)
30
µ=6
25
Exact
ATC−40, κ=1
ATC−40
20
D, cm
15
10
5
EC
0
0 0.5 1 1.5 2 2.5 3
Tn, s
(b)
Figure 12. Comparison of deformations due to El Centro (1940) ground motion from
approximate procedure and nonlinear response history analysis: (a) µ = 2, and (b) µ = 6.
30
100
µ=2
ATC−40, κ=1
ATC−40
50
Error (%)
−50
EC
−100
0 0.5 1 1.5 2 2.5 3
Tn, s
(a)
100
µ=6
ATC−40, κ=1
ATC−40
50
Error (%)
−50
EC
−100
0 0.5 1 1.5 2 2.5 3
Tn, s
(b)
Figure 13. Error in deformations due to El Centro (1940) ground motion computed by
approximate procedure: (a) µ = 2, and (b) µ = 6.
31
El Centro, Imperial Valley Earthquake (1940) Corralitos, Loma Prieta Earthquake (1989)
200 200
µ=6 µ=6
150 150
µ=2 µ=2
100 100
Error (%)
Error (%)
50 50
0 0
−50 −50
−100 −100
0 0.5 1 T , s 1.5 2 2.5 3 0 0.5 1 T , s 1.5 2 2.5 3
n n
(a) (b)
Sylmar County Hospital, Northridge Earthquake (1994) Pacoima Dam, San Fernando Earthquake (1971)
200 200
µ=6 µ=6
150 150
µ=2 µ=2
100 100
Error (%)
Error (%)
50 50
0 0
−50 −50
−100 −100
0 0.5 1 T , s 1.5 2 2.5 3 0 0.5 1 T , s 1.5 2 2.5 3
n n
(c) (d)
Lucerne Valley, Landers Earthquake (1992) SCT, Mexico City Earthquake (1985)
200 200
µ=6 µ=6
150 150
µ=2 µ=2
100 100
Error (%)
Error (%)
50 50
0 0
−50 −50
−100 −100
0 0.5 1 T , s 1.5 2 2.5 3 0 0.5 1 T , s 1.5 2 2.5 3
n n
(e) (f)
Figure 14. Error in deformations computed by approximate procedure for six ground
motions.
32
DESIGN SPECTRUM
The ATC-40 Procedure B is implemented for a wide range of T n and µ values with the
excitation characterized by the elastic design spectrum of Fig. 7. The yield strength was defined
by Eq. (8) with A y determined from the inelastic design spectrum corresponding to the selected
ductility factor. The resulting approximate values of deformations will be compared in this
section with those determined directly from the design spectrum, as described next.
Given the properties T n , ζ, f y and α of the bilinear hysteretic system and the elastic
design spectrum, the earthquake-induced deformation of the system can be determined directly
from the design spectrum. The peak deformation D of this system is given by
D = µ Dy (9)
with the yield deformation defined by
2
 Tn  (10)
D y =   Ay
 2π 
where Ay is the pseudo-acceleration related to the yield strength, f y , by Eq. (8). Putting Eqs.
(9) and (10) together gives
2
  (11)
D = µ  T n  Ay
 2π 
The yield strength reduction factor is given by
fo A (12)
Ry = =
f y Ay
where
 A (13)
f o =   w
g
is the minimum yield strength required for the structure to remain elastic; A is the pseudo-
acceleration ordinate of the elastic design spectrum at (T n , ζ ) . Substituting Eq. (12) in Eq. (11)
gives
1  
2 (14)
D = µ  Tn  A
R y  2π 
Equation (14) provides a convenient way to determine the deformation of the inelastic system
from the design spectrum. All that remains to be done is to determine µ for a given R y ; the latter
is known from Eq. (12) for a structure with known f y .
33
Presented in Fig. 15 are the deformations determined by Eq. 14 using three different
R y − µ − T n equations: Newmark and Hall (1982); Krawinkler and Nassar (1992) for
elastoplastic systems; and Vidic, Fajfar and Fischinger (1994) for bilinear systems. The
equations describing these relationships are presented later in this report. Observe that the three
recommendations lead to similar results except for T n < 0.3 sec , indicating that the inelastic
design spectrum is a reliable approach to estimate the earthquake-induced deformation of
yielding systems.
500 Acceleration Velocity Displacement

100
50
20
D, cm
10 NH
5 KN
VFF
2
1
0.5
0.2 µ=4
0.1
0.05 0.1 0.2 0.5 1 2 5 10 20 50
Tn, s
Figure 15. Deformation of inelastic systems (µ= 4) determined from inelastic design spectra
using three R y − µ − T n equations: Newmark-Hall (NH), Krawinkler-Nassar (KN), and
Vidic-Fajfar-Fischinger (VFF).
The deformation estimates by the ATC-40 method are compared in Fig. 16 with those
from inelastic design spectra presented in Fig. 15. Relative to these “reference” values, the
percentage discrepancy in the approximate result is plotted in Fig. 17. The results of Figs. 16 and
17 permit the following observations. The approximate procedure leads to significant
discrepancy, except for very long periods ( T n > T f in Fig. 7). The magnitude of this discrepancy
depends on the design ductility and the period region. In the acceleration-sensitive (T n < T c ) and
displacement-sensitive (T d <T n < T f ) regions (Fig. 7), the approximate procedure significantly
underestimates the deformation; the discrepancy increases with increasing µ. In the velocity-
sensitive (T c <T n < T d ) region, the ATC-40 procedure significantly underestimates the
deformation for µ = 2 and 4, but overestimates it for µ = 8 and is coincidentally accurate for
µ = 6.
34
In passing, note that the ATC-40 procedure is deficient relative to even the elastic design
spectrum in the velocity-sensitive and displacement-sensitive regions (T n > T c ) . For T n in these
regions, the peak deformation of an inelastic system can be estimated from the elastic design
spectrum, using the well-known equal-displacement rule (Veletsos and Newmark, 1960).
However, the ATC-40 procedure requires analyses of several equivalent linear systems and still
produces worse results.
35
100
50
20
D, cm
10
5 Design Spectrum (NH)
ATC−40
2
1
0.5
0.2 µ=4
0.1
0.05 0.1 0.2 0.5 1 2 5 10 20 50
Tn, s
(a)
100
50
20
D, cm
10
5 Design Spectrum (KN)
ATC−40
2
1
0.5
0.2 µ=4
0.1
0.05 0.1 0.2 0.5 1 2 5 10 20 50
T ,s
n
(b)
100
50
20
D, cm
10
5 Design Spectrum (VFF)
ATC−40
2
1
0.5
0.2 µ=4
0.1
0.05 0.1 0.2 0.5 1 2 5 10 20 50
Tn, s
(c)
Figure 16. Comparison of deformations computed by ATC-40 procedure with those from
three different inelastic design spectra (µ = 4): (a) Newmark and Hall (1982), (b)
Krawinkler and Nassar (1992), and (c) Vidic, Fajfar and Fischinger (1994).
36
100
Sensitive Sensitive Sensitive
50 µ=2
Discrepancy (%)
µ=4
µ=6
µ=8
0
−50
NH
−100
0.05 0.1 0.2 0.5 1 2 5 10 20 50
T ,s
n
(a)
100
50 µ=2
Discrepancy (%)
µ=4
µ=6
µ=8
0
−50
KN
−100
0.05 0.1 0.2 0.5 1 2 5 10 20 50
T ,s
n
(b)
100
50 µ=2
Discrepancy (%)
µ=4
µ=6
µ=8
0
−50
VFF
−100
0.05 0.1 0.2 0.5 1 2 5 10 20 50
T ,s
n
(c)
Figure 17. Discrepancy in deformations computed by ATC-40 procedure relative to three

different inelastic design spectra: (a) Newmark and Hall (1982), (b) Krawinkler and Nassar
(1992), and (c) Vidic, Fajfar and Fischinger (1994).
37
IMPROVED PROCEDURES
Presented next are two improved procedures that eliminate the errors (or discrepancies) in
the ATC-40 procedures, but retain their graphical appeal. Procedures A and B that are presented
are akin to ATC-40 Procedures A and B, respectively. The improved procedures use the well-
known constant-ductility design spectrum for the demand diagram, instead of the elastic design
spectrum for equivalent linear systems in ATC-40 procedures.
INELASTIC DESIGN SPECTRUM
A constant-ductility design spectrum is established by reducing the elastic design

spectrum by appropriate ductility-dependent factors that depend on T n . The earliest
recommendation for the reduction factor, R y (Eq. 12), goes back to the work of Veletsos and
Newmark (1960), which is the basis for the inelastic design spectra developed by Newmark and
Hall (1982). Starting with the elastic design spectrum of Fig. 7 and these R y − µ relations for
acceleration-, velocity-, and displacement-sensitive spectral regions, the inelastic design
spectrum constructed by the procedure described in Chopra (1995, Section 7.10) is shown in Fig.
18a.
In recent years, several recommendations for the reduction factor have been developed
(Krawinkler and Nassar, 1992; Vidic, Fajfar, and Fischinger, 1994; Riddell, Hidalgo, and Cruz,
1989; Tso and Naumoski, 1991; Miranda and Bertero, 1994). Based on two of these
recommendations, the inelastic design spectrum is shown in Figs. 18b and 18c. For a fixed µ = 2,
the inelastic spectra from Fig. 18 are compared in Fig. 19. The three spectra are very similar in
the velocity-sensitive region of the spectrum, but differ in the acceleration-sensitive region. An
improved procedure based on such inelastic design spectra is presented in two versions that
follow.
INELASTIC DEMAND DIAGRAM
The inelastic design spectra of Fig. 18 will be plotted in the A-D format to obtain the
corresponding demand diagrams. The peak deformation D of the inelastic system is given by Eq.
(11) where A y is known from Fig. 18 for a given T n and µ. Determined corresponding to the
three inelastic design spectra in Fig. 18, such data pairs ( A y ,D) are plotted to obtain the demand
diagram for inelastic systems (Fig. 20).
38
3
NH
2.5
2 µ=1
Ay, (g)
1.5
1.5
1 2
0.5 4
6
8
0
0 1 2 3 4 5
Tn, s
(a)
3
KN
2.5
2 µ=1
Ay, (g)
1.5
1.5
1 2
0.5 4
6
8
0
0 1 2 3 4 5
Tn, s
(b)
3
VFF
2.5
2 µ=1
Ay, (g)
1.5
1.5
1
2
0.5 4
6
8
0
0 1 2 3 4 5
Tn, s
(c)
Figure 18. Inelastic design spectra: (a) Newmark and Hall (1982), (b) Krawinkler and
Nassar (1992), and (c) Vidic, Fajfar and Fischinger (1994).
39
2.5
µ=2
1.5
NH
A ,g
KN
y
VFF
1
0.5
0
0 1 2 3 4 5
Tn, s
Figure 19. Pseudo-acceleration design spectrum for inelastic systems (µ= 2) using three
R y − µ − T n equations: Newmark-Hall (NH), Krawinkler-Nassar (KN), and Vidic-Fajfar-
Fischinger (VFF).
PROCEDURE A
This procedure, which uses the demand diagram for inelastic systems (Fig. 20), will be
illustrated with reference to six elastoplastic systems defined by two values of T n = 0.5 and 1.0
sec and three different yield strengths, given by Eq. (8) corresponding to µ = 2, 4, and 6,
respectively. Superimposed on the demand diagrams are the capacity diagrams for three inelastic
systems with T n = 0.5 sec (Figs. 21a, 22a, and 23a) and T n = 1.0 sec (Figs. 21b, 22b, and 23b).
The yielding branch of the capacity diagram intersects the demand diagram for several µ values.
One of these intersection points, which remains to be determined, will provide the deformation
demand. At the one relevant intersection point, the ductility factor calculated from the capacity
diagram should match the ductility value associated with the intersecting demand curve.
Determined according to this criterion, the deformation for each system is noted in Figs. 21 to
23. Implementation of this procedure is illustrated for two systems.
Examples
The yield deformation of System 1 is u y = 3.724 cm . The yielding branch of the capacity
diagram intersects the demand curves for µ = 1, 2, 4, 6, and 8 at 133.93 cm, 66.96 cm, 33.48 cm,
22.3 cm, and 16.5 cm, respectively (Fig. 21a). Dividing by u y , the corresponding ductility
factors are 133.93÷3.724=35.96 (which exceeds µ = 1 for this demand curve),
66.96÷3.724=17.98 (which exceeds µ = 2 for this demand curve), 33.48÷3.724=8.99 (which
exceeds µ = 4 for this demand curve), 22.3÷3.724=6 (which matches µ = 6 for this demand
40
curve), and 16.5÷3.724=4.43 (which is smaller than µ = 8 for this demand curve). Thus, the
ductility demand is 6 and the deformation of System 1 is D = 22.3 cm.
For System 3, u y = 9.681cm . The yielding branch of the capacity diagram intersects the
demand curve for µ = 1 at 51.34 cm (Fig. 21a). The corresponding ductility factor is
51.34÷9.681=5.3, which is larger than the µ = 1 for this demand curve. The yielding branch of
the capacity diagram also intersects the demand curve for µ = 2 continuously from 9.681 cm to
25.2 cm, which correspond to ductility factors of 1 to 2.6. The intersection point at 19.29 cm
corresponds to ductility factor = 19.39÷9.681=2 which matches µ = 2 for this demand curve.
Thus, the ductility demand is 2 and the deformation of System 3 is D = 19.39 cm.
41
0.5 s 1s
3
NH
µ=1
2 2s
1.5
Ay, g
2
1
4
6
8
5s
10 s
0
0 50 100 150 200
D, cm
(a)
0.5 s 1s
3
KN
µ=1
2 2s
Ay, g
1.5
2.0
1
4.0
6.0
5s
8.0
10 s
0
0 50 100 150 200
D, cm
(b)
0.5 s 1s
3
VFF
µ=1
2 2s
Ay, g
1.5
2
1
4 5s
8 6
10 s
0
0 50 100 150 200
D, cm
(c)
Figure 20. Inelastic demand diagrams: (a) Newmark and Hall (1982), (b) Krawinkler and
Nassar (1992), and (c) Vidic, Fajfar and Fischinger (1994).
42
0.5 s 1s
3
NH
µ=1
2 2s
(19.39,1.562)
Ay, g
2
1 (22.32,0.8992)
4
6
8 (22.32,0.5995)
5s
10 s
0
0 50 100 150 200
D, cm
(a)
0.5 s 1s
3
NH
µ=1
2 2s
Ay, g
2
1 (44.64,0.8992)
4
6 (44.64,0.4496)
8
5s
(44.64,0.2997)
10 s
0
0 50 100 150 200
D, cm
(b)
Figure 21. Application of improved Procedure A using Newmark-Hall (1982) inelastic
design spectrum: (a) Systems 1 to 3, and (b) Systems 4 to 6.
43
0.5 s 1s
3
KN
µ=1
2 2s
(17.2,1.562)
Ay, g
1.77
1 (18.15,0.8992)
3.25
(19.11,0.5995)
5.14 5s
10 s
0
0 50 100 150 200
D, cm
(a)
0.5 s 1s
3
KN
µ=1
2 2s
Ay, g
1.97
1 (43.97,0.8992)
3.80
(42.46,0.4496)
5.56 5s
(41.37,0.2997)
10 s
0
0 50 100 150 200
D, cm
(b)
Figure 22. Application of improved Procedure A using Krawinkler-Nassar (1992) inelastic
44
0.5 s 1s
3
VFF
µ=1
2 2s
(16.54,1.562)
Ay, g
1.71
1 (17.02,0.8992)
3.05
(17.43,0.5995)
4.69 5s
10 s
0
0 50 100 150 200
D, cm
(a)
0.5 s 1s
3
VFF
µ=1
2 2s
Ay, g
1.73
1 (38.58,0.8992)
3.32
(37,0.4496)
4.97 5s
(36.94,0.2997)
10 s
0
0 50 100 150 200
D, cm
(b)
Figure 23. Application of improved Procedure A using Vidic-Fajfar-Fischinger (1994)
inelastic design spectrum: (a) Systems 1 to 3, and (b) Systems 4 to 6.
45
Observe that for the presented examples, the ductility factor at the intersection point
matched exactly the ductility value associated with one of the demand curves because the f y
values were chosen consistent with the same µ values for which the demand curves have been
plotted. In general this is not the case and interpolation between demand curves for two • values
would be necessary. Alternatively, the demand curves may be plotted at a finer • interval
avoiding the need for interpolation.
Comparison with ATC-40 Procedure A
The improved procedure just presented gives the deformation value consistent with the
selected inelastic design spectrum (Table 15), while retaining the attraction of graphical
implementation of the ATC-40 Procedure A. Comparison of Figs. 21 (or 22 or 23) and 5
indicates that the two procedures are similar in the sense that the desired deformation is
determined at the intersection of the capacity diagram and the demand diagram. However, the
two procedures differ fundamentally in an important sense; the demand diagram used is
different: the constant-ductility demand diagram for inelastic systems in the improved procedure
(Figs. 21 to 23) versus the elastic demand diagram in ATC-40 Procedure A for equivalent linear
systems (Fig. 5).
Observe that equivalent linear systems are analyzed using the elastic design spectrum for
a range of damping values, wide enough to cover the large damping expected for equivalent
linear systems (Fig. 3). However, most existing rules for constructing elastic design spectra are
limited to ζ = 0 to 20% (Chopra, 1995, Section 6.9).
PROCEDURE B
This version of the improved procedure avoids construction of the inelastic design
spectrum. The peak deformation D of an inelastic system with properties T n , ζ, and f y is
determined by the following sequence of steps:
1. Plot the capacity diagram and the 5%-damped elastic demand diagram of Fig. 7 in A-D
format.
2. Assume the expected ductility demand µ; start with µ = 1 .
3. Determine A y (T n , ζ, µ) from the inelastic design spectrum for the estimated µ and calculate
D from Eq. (14).
4. Plot the point with coordinates D and Ay .
5. Check if the curve generated by connecting similar points intersects the capacity diagram. If
not, repeat Steps 3 and 4 with larger values of µ; otherwise go to Step 6.
6. The earthquake-induced deformation demand D is given by the D-value at the intersection
point.
46
Examples
This procedure is implemented for Systems 1 to 6 (Table 5) with the earthquake
excitation characterized by the elastic design spectrum of Fig. 7. The results are summarized in
Table 15; intermediate results are available in Tables 16 and 17. The inelastic design spectrum of
Newmark and Hall (1982) provides the D y , A y pairs for T n = 0.5 sec and 1.0 sec in Tables 16
and 17, respectively; and D is determined by Step 3. The (D, A y ) pairs are plotted to obtain the
curve A-B in Figs. 24a and 24b. The 5%-damped elastic demand diagram and capacity diagrams
for the selected systems are also shown; however, a plot of the elastic demand diagram is not
essential to the procedure. The intersection point between the curve A-B and the capacity
diagram gives the system deformation: D = 22.32 cm , D = 22.32 cm and D = 19.39 cm for
Systems 1, 2 , and 3, respectively (Fig. 24a) and D = 44.64 cm for Systems 4 to 6 (Fig. 24b). In
the latter case, the deformation of the inelastic system is independent of the yield strength and
equals that of the corresponding linear system because T n is in the velocity-sensitive spectral
region. This is the well-known equal displacement rule.
Comparison with ATC-40 Procedure B
The improved procedure just presented gives the deformation value consistent with the
inelastic design spectrum, while retaining the attraction of a graphical implementation of ATC-
40-Procedure B. Comparison of Figs. 24 and 11 indicates that the two procedures are graphically
similar. However, they differ fundamentally in one important sense. Each point on the curve A-B
(Fig. 24) in the improved procedure is determined by analyzing an inelastic system. In contrast
the ATC-40-Procedure B gives a point on the curve A-B (Fig. 11) by analyzing an equivalent
linear system.
ALTERNATIVE DEFINITION OF EQUIVALENT DAMPING
We digress briefly to observe that the capacity spectrum method based on the elastic
design spectrum has been modified to use an alternative definition of equivalent viscous
damping, ζeq (Freeman, 1998; WJE, 1996). This ζeq is derived by equating the peak
deformation of the equivalent linear system, determined from the elastic design spectrum
(Chopra, 1995; Section 6.9), to the peak deformation of the yielding system, determined from the
inelastic design spectrum (Chopra, 1995; Section 7.10). The capacity spectrum method,
modified in this way, should give essentially the same deformation as the improved procedure.
However, we see little benefit in making this detour when the well-known constant-ductility
inelastic design spectra can be used directly in the improved procedure.
47
Table 15. Results from Improved Procedure analysis of six systems for design spectrum.
System Dimproved D spectrum Discrepancy

(cm) (cm) (%)
1 22.32 22.32 0
2 22.32 22.32 0
3 19.39 19.39 0
4 44.64 44.64 0
5 44.64 44.64 0
6 44.64 44.64 0
Table 16. Detailed results from improved Procedure B analysis of Systems 1 to 3 for design
spectrum.
µ Dy Ay D
(cm) (g) (cm)
1 16.794 2.706 16.794
1.25 13.713 2.21 17.141
1.5 11.875 1.914 17.813
1.75 10.622 1.712 18.588
2 9.696 1.562 19.392
2.25 8.977 1.447 20.198
2.5 8.397 1.353 20.993
2.75 7.917 1.276 21.772
3 7.44 1.199 22.321
3.25 6.868 1.107 22.321
3.5 6.377 1.028 22.321
3.75 5.952 0.959 22.321
4 5.58 0.8992 22.321
4.25 5.252 0.846 22.321
4.5 4.96 0.799 22.321
4.75 4.699 0.757 22.321
5 4.464 0.719 22.321
5.25 4.252 0.685 22.321
5.5 4.058 0.654 22.321
5.75 3.882 0.626 22.321
6 3.72 0.5995 22.321
6.25 3.571 0.575 22.321
6.5 3.434 0.553 22.321
6.75 3.307 0.533 22.321
7 3.189 0.514 22.321
48
Table 17. Detailed results from improved Procedure B analysis of Systems 4 to 6 for design
spectrum.
µ Dy Ay D
(cm) (g) (cm)
1 44.642 1.798 44.642
1.25 35.714 1.439 44.642
1.5 29.761 1.199 44.642
1.75 25.51 1.028 44.642
2 22.321 0.8992 44.642
2.25 19.841 0.799 44.642
2.5 17.857 0.719 44.642
2.75 16.233 0.654 44.642
3 14.881 0.599 44.642
3.25 13.736 0.553 44.642
3.5 12.755 0.514 44.642
3.75 11.905 0.48 44.642
4 11.161 0.4496 44.642
4.25 10.504 0.423 44.642
4.5 9.92 0.4 44.642
4.75 9.398 0.379 44.642
5 8.928 0.36 44.642
5.25 8.503 0.343 44.642
5.5 8.117 0.327 44.642
5.75 7.764 0.313 44.642
6 7.44 0.2997 44.642
6.25 7.143 0.288 44.642
6.5 6.868 0.277 44.642
6.75 6.614 0.266 44.642
7 6.377 0.257 44.642
49
0.5 s 1s
3
A NH
2 2s
(19.39,1.562)
Ay, g
1 (22.32,0.8992)
(22.32,0.5995)
5s
B
10 s
0
0 50 100 150 200
D, cm
(a)
0.5 s 1s
3
NH
2 2s
A
Ay, g
1 (44.64,0.8992)
(44.64,0.4496)
5s
(44.64,0.2997)
B 10 s
0
0 50 100 150 200
D, cm
(b)
Figure 24. Application of improved Procedure B using Newmark-Hall (1982) inelastic
50
IMPROVED PROCEDURE: NUMERICAL VERSION
BASIC CONCEPT
The improved procedures presented in the preceding section were implemented

graphically, in part, to highlight the similarities and differences relative to the Nonlinear Static
Procedure in the ATC-40 report. The graphical implementation of the first version of the
improved procedure is especially attractive as the desired earthquake-induced deformation is
determined at the intersection of the capacity and demand diagrams. However, the graphical
feature is not essential and the procedure can be implemented numerically. Such a procedure
using R y − µ − T n equations is presented in this section.
Ry – µ −Tn EQUATIONS
The R y − µ − T n equations for elastoplastic systems, consistent with the Newmark-Hall

inelastic design spectra are (Chopra, 1995; Section 7.10):
1 Tn <Ta
(2µ −1)β / 2 Ta < Tn < Tb

 2µ − 1 T b < T n < T c'
=
Ry  (15a)
T n µ T c' < T n < T c
T c
µ Tn >Tc
where
β = ln (T n T a ) ln (T b T a ) (15b)
and the T a , T b , and T c are defined in Fig. 7 and T c' is the period where the constant-A and
constant-V branches of the inelastic design spectrum intersect (Chopra, 1995, Section 7.10).
Recasting Eq. (15) gives µ as a function of Ry :
 Undefined Tn <Ta
(1 + R 2y / β ) 2 Ta < Tn < Tb
(
 1 + R 2y ) 2 T b < T n < T c'
µ= (16)
T c R y T c' < T n < T c
T n
 R y Tn >Tc
For a given R y , µ can be calculated for all T n except for T b < T n < T c , wherein two possibilities
need to be checked since T c' itself depends on µ (see Appendix B).
Based on the earthquake response of bilinear systems, Krawinkler and Nassar (1992)
have developed the following R y − µ − T n equations:
51
R y = [c(µ −1)+1] (17)
1/ c
where
a (18)
c (T n , α) = T n a +
b
1+ Tn Tn
and the numerical coefficients depend on the slope αk of the yielding branch (Fig. 2a): a = 1
and b = 0.42 for α = 0% ; a = 1 and b = 0.37 for α = 2% ; a = 0.8 and b = 0.29 for α = 10% .
Recasting Eq. (18) provides µ as a function of R y :
µ = 1+ (R y − 1)
1 c (19)
c
For given values of R y and α, µ can be calculated from Eq. (19).
Based on the earthquake response of bilinear systems, Vidic, Fajfar and Fischinger
(1994) have developed the following R y − µ − T n equations:

1.35 (µ −1)
0.95 T n
+1 T n ≤ To
=
Ry  To (20)
1.35 (µ −1)0.95 + 1 Tn > To
where
T o = 0.75µ T c ≤ T c (21)
0.2
Recasting Eq. (20) gives µ as a function of R y :
  o
1.053
1 + 0.74(R y −1) 
T
µ =   Tn ≤ To (22)
Tn
1 + [0.74( −1)]1.053
 Ry Tn > To
Since T o in Eq. (22) depends on µ (Eq. 21), the value of µ corresponding to a given R y is
determined by solving a nonlinear equation iteratively unless the simpler relation, T o = T c , is
assumed.
Figure 25 shows plots of R y v’s T n for selected values of µ based on Eqs. (15), (17) for
α = 0, and (20). In Fig. 26, µ is plotted against T n from Eqs. (16), (19) for α = 0, and (22).
Observe the similarity among the three sets of results, indicating consensus among different
researchers.
The peak deformation of systems 1 to 6 (Table 5) are determined from Eq. (14) using
Ry − µ − T n relations of Eq. (16), (19), and (22). Detailed calculations are presented in Appendix
B and the results are summarized in Table 18. Observe that the deformation values computed
using R y − µ − T n equations are identical to those determined by the graphical procedure (Figs.
21 to 23) except for round-off differences.
52
10
NH
KN
VFF
8
R
y 6 µ=6
2
µ=2
1
0
0 0.5 1 1.5 2 2.5 3
Tn,s
Figure 25. Variation of R y with T n for selected ductility values based on three different
sources: Newmark and Hall (1982), Krawinkler and Nassar (1992), and Vidic, Fajfar, and
Fischinger (1994).
20
16 NH
KN
VFF
12
µ
8
Ry = 6
4
Ry = 2
0
0 0.5 1 1.5 2 2.5 3
Tn,s
Figure 26. Variation of µ with T n for selected R y values based on three different sources:
Newmark and Hall (1982), Krawinkler and Nassar (1992), and Vidic, Fajfar, and
Fischinger (1994).
53
Table 18. Results from numerical implementation of improved procedure using three R y − µ − T n equations.
System Properties Newmark-Hall Krawinkler-Nassar Vidic et al.

System Tn A Ay Dy = uy Ry µ D µ D µ D
( f y ÷ w) ( A ÷ Ay ) (Eq. 16) (µ × D y) (Eq. 19) (µ × D y) (Eq. 22) (µ × D y)
(s) (g) (cm) (cm) (cm) (cm)
(g)
1 0.5 2.7062 0.5995 3.7202 4.51 5.99 22.29 5.14 19.11 4.69 17.43
2 0.8992 5.5803 3.01 3.99 22.29 3.25 18.15 3.05 17.02
3 1.5624 9.6962 1.73 2.00 19.39 1.77 17.20 1.71 16.54
4 1 1.7984 0.2997 7.4403 6.00 6.00 44.64 5.56 41.37 4.97 36.94
5 0.4496 11.160 4.00 4.00 44.64 3.80 42.46 3.32 37.00
6 0.8992 22.321 2.00 2.00 44.64 1.97 43.97 1.73 38.58
54
CONSISTENT TERMINOLOGY
Many new terms that have been introduced in connection with simplified analysis of
inelastic systems are examined in this section and, where necessary, better terminology is
recommended:
1. Demand Spectrum. The term “spectrum” has traditionally implied a function of frequency or
period. For example, response spectrum is a plot of the peak value of a response quantity as a
function of the natural vibration period (or frequency) of the SDF system. Another example:
Fourier Spectrum of ground acceleration is a plot of the amplitude of the Fourier transform
of the excitation against exciting frequency. The “Response Spectrum” terminology was
introduced in the 1930s within the context of earthquake engineering, whereas the “Fourier
Spectrum” terminology has existed for much longer. Given this background, “spectrum” is
inappropriate to describe a plot of pseudo-acceleration v's deformation. The terminology
Demand Diagram has therefore been used in this investigation.
2. Capacity Spectrum. For the same reasons, the recommended terminology is Capacity
Diagram.
3. Acceleration-Displacement Response Spectrum (ADRS) Format. For the same reasons, the
recommended terminology is A-D format.
4. Nonlinear Static Procedure (NSP). The capacity diagram is determined by nonlinear static
analysis of the structure, but the demand diagram is determined by dynamic analysis.
Because the NSP involves use of both diagrams, the NSP terminology is inappropriate and
the suggested alternative is Simplified Dynamic Analysis.
5. Modal Participation Factor. This traditional terminology for Γn (Eq. 1) implies that it is a
measure of the degree to which the nth mode participates in the response. However, this is
misleading because Γn is not independent of how the mode is normalized, nor a measure of
the modal contribution to a response quantity (Chopra, 1995; section 13.1).
CONCLUSIONS
This investigation of capacity-demand-diagram methods to estimate the earthquake-

induced deformation of inelastic SDF systems has led to the following conclusions:
1. Based on the belief that the deformation of an inelastic system can be estimated by an
iterative method requiring analysis of a sequence of equivalent linear systems, the ATC-40
Procedure A did not converge for some of the systems analyzed. It converged in many cases
but not to the exact deformation determined by nonlinear response history analysis of the
inelastic system, nor to the value determined from the inelastic design spectrum. Thus,
convergence of this iterative procedure is deceptive because it can leave the erroneous
impression that the calculated deformation is accurate. This approximate procedure therefore
does not meet the basic requirement of a rational iterative procedure: it should always
converge to the “exact” result after a sufficient number of iterations.
55
2. The ATC-40 Procedure B always gives a unique value of deformation, the same as
determined by Procedure A if it converged.
3. The peak deformation of inelastic systems, determined by ATC-40 procedures, when
compared against results of nonlinear response history analysis for several ground motions
were shown to be inaccurate. The approximate procedure underestimates significantly the
deformation for a wide range of T n values with errors approaching 50%, implying that the
estimated deformation is only about half of the “exact” value.
4. The damping modification factor, κ, in ATC-40 procedures improves the deformation
estimate only marginally. Therefore the κ factor is not attractive, especially because it is
based primarily on judgement.
5. The ATC-40 procedures were implemented for a wide range of T n and µ values with the
excitation characterized by an elastic design spectrum. The resulting estimate of deformation
for the inelastic system was compared with the deformation determined from the inelastic
design spectrum using three different R y − µ − T n equations (Newmark and Hall, 1982;
Krawinkler and Nassar, 1992; Vidic, Fajfar, and Fischinger, 1994), all of which provided
similar results. Relative to these “reference” values, the approximate procedure significantly
underestimates the deformation for a wide range of T n and µ values.
6. The ATC-40 procedures are deficient relative to even the elastic design spectrum in the
velocity-sensitive and displacement-sensitive regions of the spectrum. For T n in these
regions, the peak deformation of an inelastic system can be estimated from the elastic design
spectrum using the well-known equal displacement rule. However, the approximate
procedure requires analyses of several equivalent linear systems and still produces worse
results.
7. An improved capacity-demand-diagram method that uses the well-known constant-ductility
design spectrum for the demand diagram has been developed and illustrated by examples.
When both capacity and demand diagrams are plotted in the A-D format, the yielding branch
of the capacity diagram intersects the demand curves for several µ values. The deformation is
given by the one intersection point where the ductility factor calculated from the capacity
diagram matches the value associated with the intersecting demand curve. This method gives
the deformation value consistent with the selected inelastic design spectrum, while retaining
the attraction of graphical implementation of the ATC-40 methods.
8. One version of the improved method is graphically similar to ATC-40 Procedure A.
However, the two differ fundamentally in an important sense; the demand diagram used is
different: the constant-ductility demand diagram for inelastic systems in the improved
procedure versus the elastic demand diagram in Procedure A for equivalent linear systems.
9. A second version of the improved method is graphically similar to ATC-40 Procedure B.
However the two differ fundamentally in one important sense. Each point on curve A-B is
determined by analyzing an inelastic system in the improved procedure (Fig. 24) but an
equivalent linear system in Procedure B (Fig. 12).
10. The improved method can be conveniently implemented numerically if its graphical features
are not important to the user. Such a procedure, based on equations relating R y and µ for
56
different T n ranges, has been presented. It has been illustrated by examples using three
different R y − µ − T n relations (Newmark and Hall, 1982; Krawinkler and Nassar, 1992;
Vidic, Fajfar, and Fischinger, 1994). The graphical and numerical implementations of the
improved method are shown to give essentially identical values for deformation.
11. The new terminology that has been introduced in recent years in connection with simplified
analysis of inelastic systems has been examined and, where necessary, better terminology
recommended:
(a) The term “spectrum” implies a function of frequency or period (e.g., response spectrum
or Fourier spectrum) and is therefore inappropriate to describe a plot of pseudo-
acceleration versus deformation. The recommended terminology is Demand Diagram and
Capacity Diagram instead of Demand Spectrum and Capacity Spectrum.
(b) Acceleration-Displacement Response Spectrum (ADRS) Format is inappropriate for the
same reason and A-D Format is preferable.
(c) Nonlinear Static Procedure (NSP) is a misleading term because the NSP uses a demand
diagram determined by dynamic analysis. The suggested alternative is Simplified
Dynamic Analysis.
57
NOTATION
α strain hardening ratio

a constant used in Krawinkler and Nassar R y − µ − T n equations
β constant used in Newmark and Hall R y − µ − T n equations
A pseudo-acceleration spectrum ordinate
Ai (2π /T n )2 D i , pseudo-acceleration corresponding to Di
Aj (2π /T n )2 D j , pseudo-acceleration corresponding to Dj
Ay (2π /T n )2 D y , pseudo-acceleration corresponding to yield deformation Dy
b constant used in Krawinkler and Nassar R y − µ − T n equations
c variable used in Krawinkler and Nassar R y − µ − T n equations
D deformation spectrum ordinate

D approx peak deformation computed from approximate procedure
D exact exact peak deformation determined by nonlinear response history analysis
D spectrum exact peak deformation determined by inelastic design spectrum
Di deformation estimate at the beginning of an iteration
Dj deformation estimate at the end of an iteration
Dy yield deformation = u y
ED energy dissipated in inelastic system (= area of hysteresis loop)

ES strain energy of the equivalent linear system
φ1 fundamental mode
φ j1 jth floor element of the fundamental mode φ1
fo minimum strength required for a system to remain elastic
f y
yield strength
g acceleration due to gravity
Γ1 fundamental mode factor defined by Eq. (1)
k initial elastic stiffness of the inelastic system
k sec secant stiffness of the equivalent linear system
58
κ damping modification factor specified in ATC-40
m mass of the system
mj lumped mass at the jth floor level
*
M1 effective modal mass for the fundamental vibration mode
N number of floors
µ ductility factor
Ry yield reduction factor
T a ,T b ,T c ,T c' , periods that define spectral regions
T d ,T e ,T f
To transition period used in Vidic, Fajfar, and Fischinger R y − µ − T n equations
T eq equivalent vibration period
Tn natural vibration period
u deformation
u go , u go , ugo peak ground displacement, velocity, and acceleration
um peak deformation of inelastic system
uy yield displacement
Vy pseudo-velocity corresponding to yield deformation D y
w weight of the system
ζ viscous damping ratio of linearly elastic system
ζ eq equivalent viscous damping ratio
ζ̂ eq viscous damping used in equivalent linear procedures
59
REFERENCES
Applied Technology Council (1996). Seismic evaluation and retrofit of concrete buildings. Report
ATC 40. November.
Bertero, V. V. (1995). Tri-service manual methods. In Vision 2000, Part 2, App. J. Sacramento,
Calif.: Structural Engineers Assn. of California.
Chopra, A. K. (1995). Dynamics of structures: theory and applications to earthquake engineering,
Chaps. 3, 6, 7, and 19. Englewood Cliffs, N.J.: Prentice Hall.
Deierlein, G. G., and Hsieh, S-H. (1990). Seismic response of steel frames with semi-rigid
connections using the capacity spectrum method. Proceedings of 4th U.S. National Conference on
Earthquake Engineering 2:863-72.
Fajfar, P. (1998). Capacity spectrum method based on inelastic demand spectra. IKIPR Report EE
3/98. September. Ljubljana, Slovenia: Univ. of Ljubljana.
Fajfar, P. (1999). Capacity spectrum method based on inelastic spectra. Earthquake Engineering
and Structural Dynamics (forthcoming).
FEMA (1997). NEHRP guidelines for the seismic rehabilitation of buildings, FEMA 273; and
NEHRP commentary on the guidelines for the seismic rehabilitation of buildings, FEMA 274.
October. Washington, D.C.: Federal Emergency Management Agency.
Freeman, S. A., Nicoletti, J. P., and Tyrell, J. V. (1975). Evaluations of existing buildings for
seismic risk — A case study of Puget Sound Naval Shipyard, Bremerton, Washington.
Proceedings of 1st U.S. National Conference on Earthquake Engineering, 113-22. Berkeley,
Calif.: EERI
Freeman, S. A. (1978). Prediction of response of concrete buildings to severe earthquake motion.
Publication SP-55, 589-605. Detroit, Mich.: American Concrete Inst.
Freeman, S. A. (1998). Development and use of capacity spectrum method. Proceedings of 6th U.S.
National Conference on Earthquake Engineering, Seattle. CD-ROM. Oakland, Calif.: EERI.
Gulkan, P., and Sozen, M. A. (1974). Inelastic responses of reinforced concrete structures to
earthquake motions. Journal of the American Concrete Inst. 71(12):604-10.
Hudson, D. E. (1965). Equivalent viscous friction for hysteretic systems with earthquake-like
excitations. Proceedings of Third World Conference on Earthquake Engineering II:185-202.
Iwan, W. D., and Gates, N. C. (1979a). Estimating earthquake response of simple hysteretic
structures. Journal of the Engineering Mechanics Div.. 105 (EM3):391-405. ASCE
Iwan, W. D., and Gates, N. C. (1979b). The effective period and damping of a class of hysteretic
structures. Earthquake Engineering and Structural Dynamics 7(3):199-212.
Jennings, P. C. (1968). Equivalent viscous damping for yielding structures. Journal of the
Engineering Mechanics Div. 94 (EM1):103-116. ASCE.
60
Kowalsky, M. J., Priestley, M. J. N., and Macrae, G. A. (1995). Displacement-based design of RC
bridge columns in seismic regions. Earthquake Engineering and Structural Dynamics 24(12):
1623-43.
Krawinkler, H., and Nassar, A. A. (1992). Seismic design based on ductility and cumulative
damage demands and capacities. Nonlinear Seismic Analysis and Design of Reinforced Concrete
Buildings, eds P. Fajfar and H. Krawinkler. New York: Elsevier Applied Science.
Miranda, E., and Bertero, V. V. (1994). Evaluation of strength reduction factors for earthquake-
resistant design. Earthquake Spectra 10(2): 357-79.
Moehle, J. P. (1992). Displacement-based design of R/C structures subjected to earthquakes.
Earthquake Spectra 8(3): 403-427.
Newmark, N. M., and Hall, W. J. (1982). Earthquake Spectra and Design. Berkeley, Calif.
Earthquake Engineering Research Inst.
Paret, T. F., Sasaki, K. K., Eilbekc, D. H., and Freeman, S. A. (1996). Approximate inelastic
procedures to identify failure mechanisms from higher mode effects. Paper No. 966, 11th World
Conference on Earthquake Engineering. Acapulco, Mexico.
Reinhorn, A. M., Li, C., and Constantinou, M. C. (1995). Experimental and analytical
investigations of seismic retrofit of structures with supplemental damping. Report No. NCEER-95-
0001. State Univ. of New York at Buffalo.
Reinhorn, A. M. (1997). Inelastic analysis techniques in seismic evaluations. Seismic design
methodologies for the next generation of codes, eds. P. Fajfar and H. Krawinkler, 277-87.
Rotterdam: Balkema.
Riddell, R., Hidalgo, P., Cruz, E. (1989). Response modification factors for earthquake resistant
design of short period buildings. Earthquake Spectra 5(3):571-90.
Shibata, A., and Sozen, M. A. (1976). Substitute structure method for seismic design in R/C.
Journal of the Structural Division 102(ST1):1-18. ASCE.
Tso, W. K., Naumoski, N. (1991). Period-dependent seismic force reduction factors for short-
period structures. Canadian Journal of Civil Engineering 18(4):568-74.
Tsopelas, P., Constantinou, M. C., Kircher, C. A., and Whittaker, A. S. (1997). Evaluation of
simplified methods of analysis for yielding structures. Report No. NCEER-97-0012, State Univ.
of New York at Buffalo.
Veletsos, A. S., and Newmark, N. M. (1960). Effects of inelastic behavior on the response of
simple system to earthquake motions. Proceedings of the 2nd World Conference on Earthquake
Engineering, Japan 2:895-912.
Vidic, T., Fajfar, P., and Fischinger, M. (1994). Consistent inelastic design spectra: strength and
displacement. Earthquake Engineering and Structural Dynamics 23(5) 507-521.
Wallace, J. W. (1995). Displacement-based design of R/C structural walls. Proceedings of 10th
European Conference on Earthquake Engineering 3:1539-144.
61
WJE. (1996). Seismic Dynamic Analysis for Buildings. Final Manuscript prepared for the U.S.
Army Engineering Division by Wiss, Janey, Elstner Assocs. Inc. Emeryville, Calif. (Unpublished).
62
APPENDIX A. DEFORMATION OF VERY-SHORT PERIOD SYSTEMS BY ATC-40
PROCEDURE B
Consider an elastoplastic system with the following properties: T n = 0.1 s , ζ = 5%,

f y = 0.2659 w , u y = 0.066 cm . The deformation demand of this system computed from ATC-40
Procedure B is D = 2.267 cm (Fig. A.1a) and the values of T eq = 0.5861s and ζˆ = 0.397 . The
eq
“exact” deformation of this system computed from nonlinear RHA is 0.3944 cm and the error is
100 × ( 2.267 − 0.3944) ÷ 0.3944 = 475%. Clearly ATC-40 Procedure B leads to unreasonably
large error for such a system.
Figure A.1b shows the close-up view of the intersection between curve A-B and the
capacity diagram wherein the demand diagrams for 5% and 39.7% damping are also included.
Observe that the demand diagram for 39.7% damping also passes through the intersection point
between curve A-B and the capacity diagram. This figure leads to the following observations.
The curve A-B can be divided into three distinct segments. Initially the curve is very
steep, it then becomes essentially flat, and finally it begins to move downward slowly. Since the
capacity of the selected system is lower than the flat portion of the curve A-B, the intersection
between the capacity diagram and the curve A-B does not occur till the third segment. On the
39.7% demand curve, the intersection occurs past the constant pseudo-acceleration region of the
spectrum. This implies that even though the initial elastic period of the system is in the very-
short period region, the intersection would occur in the constant-V region if the capacity of the
system is smaller than the pseudo-acceleration value in the constant-A region of the response
spectrum corresponding to ζˆ eq = 0.397 . As a result, the equivalent period at this intersection
point would have to be significantly larger than the initial elastic period (for the selected system
T n = 0.1s and T eq = 0.5861s ) which in turn would lead to large deformation demand.
63
0.5 s 1s
1
0.8
5% Demand Diagram
A
A, g 0.6
2s
0.4
(2.267,0.2659)
0.2 B
5s
10 s
0
0 10 20 30 40 50
D, cm
(a)
0.1 s 0.2 s
1
5% Demand Diagram
0.8 0.5 s
A
0.6
A, g
0.4
(2.267,0.2659)
Capacity Diagram B
0.2 1s
39.7% Demand Diagram
2s
0
0 1 2 3 4 5
D, cm
(b)
Figure A.1. ATC-40 Procedure B applied to a system with T n = 0.1s and A y = 0.2659 g: (a)
deformation demand; and (b) close-up view near the intersection between the demand and
capacity diagrams.
64
APPENDIX B: EXAMPLES USING R y − µ − T n EQUATIONS
SYSTEM 1

Ay = f y ÷ w = 0.5995g . The pseudo-acceleration for linear-elastic system with T n = 0.5 s , ζ = 5%
is A = 2.7062 g . For this system, D y = u y = (2 π T n ) × A y × g = (2 π 0.5) × 0.5995 × 980 = 3.7202
2 2
cm and R y = 2.7062 ÷ 0.5995 = 4.51.
Newmark-Hall Equation (Eq. 16)

The periods corresponding to various regions of the Newmark-Hall design spectrum are
T a = 1 / 33 s , T b = 0.125 s , and T c = 0.6634 s ; T c' is unknown at this time because it depends on
µ that is yet to be determined. Since T c' is unknown, it is not clear which equation — the third or
the fourth in Eq. (16) — should be used to determine µ. Assuming that T b < T n < T c ' , gives
µ = (1 + R 2y ) ÷ 2 = (1 + 4.512 ) ÷ 2 = 10.67 . For this value of µ, T c ' = T c 2µ − 1 µ
= 0.6634 2 × 10.67 − 1 10.67 = 0.2804 s. Clearly the assumption that T b < T n < T c ' is not correct.
Therefore, we must use the available equation for T c ' < T n < T c , which gives µ = R y × T c ÷ T n
= 4.51 × 0.6634 ÷ 0.5 = 5.99 . For this value of µ, T c ' = 0.6634 2 × 5.99 − 1 5.99 = 0.3674 s and
the assumption that T b < T n < T c ' is valid. The deformation demand D = 5.99 × 3.7202 = 22.29
cm.
Krawinkler-Nassar Equation (Eq. 19)
For elastoplastic force-deformation behavior, a=1 and b=0.42. Then
c = 0.5 ÷ (1 + 0.5) + 0.42 ÷ 0.5 = 1.1733 (Eq. 18) and µ = 1 + (1 ÷ 1.1733) × (4.511.1733 − 1) = 5.14.
The deformation demand D = 5.14 × 3.7202 = 19.11 cm.
Vidic-Fajfar-Fischinger Equation (Eq. 22)
Since µ is unknown at this time, T o (Eq. 21) can not be determined at this time. Assume
that T o = T c = 0.6634 s (Fajfar 1998), T n = 0.5s < T o . The first of the two equations in Eq. (22)
then gives µ = 1 + [0.74×(0.6634 0.5)(4.51−1)] = 4.69 . The deformation demand D=
1.053
4.69 × 3.7202 = 17.43 cm.
SYSTEM 2

Ay = f y ÷ w = 0.8992 g . The pseudo-acceleration for linear-elastic system with T n = 0.5 s , ζ = 5%
65
is A = 2.7062 g . For this system, D y = u y = (2 π T n ) × A y × g = (2π 0.5) × 0.8992 × 980 = 5.5803
2 2
cm and R y = 2.7062 ÷ 0.8992 = 3.01.

Assuming that T b < T n < T c ' , gives µ = (1 + 3.012 ) ÷ 2 = 5.03 . For this value of µ, T c' =
0.6634 × 2 × 5.03 − 1 / 5.03 = 0.3970 s. Clearly the assumption that T b < T n < T c ' is not correct.
Therefore, we must use the available equation for T c ' < T n < T c , which gives µ
= 3.01 × 0.6634 ÷ 0.5 = 3.99 . For this value of µ, T c ' = 0.6634 2 × 3.99 − 1 3.99 = 0.4393 s and
the assumption that T b < T n < T c ' is valid. The deformation demand D = 3.99 × 5.5803 = 22.29
cm.

c = 0.5 ÷ (1 + 0.5) + 0.42 ÷ 0.5 = 1.1733 (Eq. 20) and µ = 1 + (1 ÷ 1.1733) × (3.011.1733
− 1) = 3.25.

Assuming that T o = T c = 0.6634 s and using the first of the two equations in Eq. (22),
µ = 1 + [0.74×(0.6634 0.5)(3.01−1)] = 3.05 .
1.053
gives The deformation demand
D = 3.05 × 5.5803 = 17.02 cm.
SYSTEM 3

Ay = f y ÷ w = 1.5624 g . The pseudo-acceleration for linear-elastic system with T n = 0.5 s , ζ = 5%
A = 2.7062 g . For this D y = (2 π 0.5) × 1.5624 × 980
2
is system, = 9.6962 cm and
R y = 2.7062 ÷ 1.5624 = 1.73.

Assuming that T b < T n < T c ' , gives µ = (1 + 1.732 ) ÷ 2 = 2.0 . For this value of µ, T c' =
0.6634 × 2 × 2.0 − 1 / 2.0 = 0.5745 s and the assumption that T b < T n < T c ' is valid. The
deformation demand D = 2.0 × 9.6962 = 19.39 cm.

c = 0.5 ÷ (1 + 0.5) + 0.42 ÷ 0.5 = 1.1733 (Eq. 20) and µ = 1 + (1 ÷ 1.1733) × (1.731.1733
− 1) = 1.77.
66
Assuming that T o = T c = 0.6634 s and using the first of the two equations in Eq. (22),
µ = 1 + [0.74×(0.6634 0.5)(1.73−1)] = 1.71 .
1.053
gives The deformation demand
D = 1.71 × 9.6962 = 16.54 cm.
SYSTEMS 4 TO 6
Consider an elastoplastic systems with the following properties: T n = 1 s , ζ = 5%,

A y = 0.2997 g , 0.4496 g , and 0.8992 g for Systems 4, 5, and 6, respectively. The pseudo-
acceleration for linear-elastic system with T n = 1 s , ζ = 5% is A = 1.7984 g . For these system,
D y = 7.4403 cm, 11.160 cm, and 22.321 cm, and corresponding values of R y = 6, 4, and 2.

For T n > T c , µ = R y = 6 for System 4, 4 for System 5, and 2 for System 6 which gives
D = µ × D y = 44.64 cm for all these examples.

c = 1.0 ÷ (1 + 1.0 ) + 0.42 ÷ 1.0 = 0.92 (Eq. 18). The values of µ for the three Systems are 5.56,
3.80, and 1.97 which give deformation demand D = 41.37 cm, 42.46 cm, and 43.97 cm.

For T n > T o , values of µ for the three Systems are 4.97, 3.32, and 1.73 which give
deformation demand D = 36.94 cm, 37.0 cm, and 38.58 cm.
67

View Content 2

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

View Content 2

Uploaded by

Copyright:

Available Formats

Pacific Earthquake Engineering

Capacity-Demand-Diagram Methods for

A report on research conducted under grant no. CMS-9812531

A report on Research Conducted Under

Report No. PEER-1999/02

EQUIVALENT LINEAR SYSTEMS.......................................................................................... 3

ATC-40 ANALYSIS PROCEDURES ......................................................................................... 7

APPENDIX B: EXAMPLES USING R y − µ − T n EQUATIONS ........................................... 65

A major challenge to performance-based seismic design and engineering of buildings is

5% Demand Demand Point

EQUIVALENT LINEAR SYSTEMS

The earthquake response of inelastic systems can be estimated by approximate analytical

For elastoplastic systems, α = 0 and Eqs. (2) and (4) reduce to

Figure 4. Variation of damping modification factor with equivalent viscous damping.

The damping modification factor, κ , based primarily on judgment, depends on the

This procedure in the ATC-40 report is described herein as a sequence of steps:

4. Compute the equivalent damping ratio ζ̂ eq from Eq. (7).

Examples: Specified Ground Motion

System Properties System Response

System Converged D approx D exact Error

Figure 5. Application of ATC-40 Procedure A to System 5 for El Centro (1940) ground

Iteration Di Ai ζ̂ eq Dj Aj Difference Comments

0.2 (11.73,0.1733) Iterations 2 & 4

Figure 6. Application of ATC-40 Procedure A to System 6 for El Centro (1940) ground

Figure 7. Newmark-Hall elastic design spectrum.

The procedure is implemented for System 5 (Table 5).

System Properties System Response

System Converged D approx D spectrum Discrepancy

39.7% Demand Diagram (1st Iteration)

Iteration Di Ai ζ̂ eq Dj Aj Difference Comments

(89.28,0.8992) Iterations 2 & 4

This procedure in the ATC-40 report is described herein as a sequence of steps:

5. Compute the peak deformation D(T eq , ζˆ eq ) and peak pseudo-acceleration A(T eq , ζˆ eq ) of an

6. Plot the point with coordinates D(T eq , ζˆ eq ) and A(T eq , ζˆ eq ) .

Examples: Specified Ground Motion

System D approx D exact Discrepancy

System D approx D spectrum Discrepancy

5% Elastic Demand Diagram

5% Elastic Demand Diagram

SPECIFIED GROUND MOTION

500 Acceleration Velocity Displacement

Figure 17. Discrepancy in deformations computed by ATC-40 procedure relative to three

INELASTIC DESIGN SPECTRUM

A constant-ductility design spectrum is established by reducing the elastic design

INELASTIC DEMAND DIAGRAM

ALTERNATIVE DEFINITION OF EQUIVALENT DAMPING

System Dimproved D spectrum Discrepancy

5% Elastic Demand Diagram

5% Elastic Demand Diagram

The improved procedures presented in the preceding section were implemented

The R y − µ − T n equations for elastoplastic systems, consistent with the Newmark-Hall

Recasting Eq. (20) gives µ as a function of R y :

System Properties Newmark-Hall Krawinkler-Nassar Vidic et al.

This investigation of capacity-demand-diagram methods to estimate the earthquake-

α strain hardening ratio

D deformation spectrum ordinate

ED energy dissipated in inelastic system (= area of hysteresis loop)

φ j1 jth floor element of the fundamental mode φ1

fo minimum strength required for a system to remain elastic

Consider an elastoplastic system with the following properties: T n = 0.1 s , ζ = 5%,

Consider an elastoplastic system with the following properties: T n = 0.5 s , ζ = 5%,

cm and R y = 2.7062 ÷ 0.5995 = 4.51.

Newmark-Hall Equation (Eq. 16)